JP6016144B2 - Transmission device, reception device, transmission method, and reception method - Google Patents

Description

  The present invention relates to a transmission device and a reception device that perform digital communication in an environment where strong external noise occurs, and a transmission method and a reception method.

  Currently, products compliant with IEEE802.11b, g, n, which is a wireless LAN (Local Area Network) standard using the 2.4 GHz band ISM (Industry-Science-Medical) band, are widely used as infrastructure for home networks. ing. The ISM band is a frequency band that is open in Japan so that it can be used without a license if it has an output of 10 mW or less. In addition to the wireless LAN standard, Zigbee (registered trademark), Bluetooth (registered trademark), cordless telephones, etc. Other communication systems also use the ISM band. Therefore, signals from other communication systems may appear as interference during wireless LAN communication. In addition, it is known that artificial noise generated by electric equipment using a high-frequency device such as a microwave oven appears in the ISM band.

  As an infrastructure of a home network other than a wireless LAN, a power line communication (PLC) system using a power line as a communication medium is known. The power line communication system is a system that performs communication by superimposing a communication signal on a long wave band or a short wave frequency band that is not used for power transmission in the power line. However, in these frequency bands, interference from communication devices that perform a plurality of power line communications with different communication methods, noise generated from electronic devices connected to the power lines, broadcast waves from broadcasting stations, external electronic devices There are leaked radio waves from.

  As a result, the communication path between the transmission apparatus and the reception apparatus of the wireless communication using the ISM band or the power line communication system is a communication path that is greatly affected by artificial noise and interference from other systems. In this document, interference from other systems, artificial noise from electronic devices, etc. are collectively expressed as external noise. In general, extraneous noise has a large power level and is characterized by appearing at a specific time and frequency. Therefore, a receiver designed for communication in an additive white gaussian noise (AWGN) environment generated by the receiver may perform optimal reception in an environment where external noise exists. The bit error rate characteristic is deteriorated. As a result, the throughput of the communication system decreases.

  In order to solve the above problem, studies have been made on optimum reception in an external noise environment such as impulsive noise. For example, Non-Patent Document 1 discloses a method of estimating necessary parameters and optimally receiving when a Middleton class A impulse radio noise model is used as a statistical model of external noise. However, the method described in Non-Patent Document 1 requires additional arithmetic processing for estimating the parameters of extraneous noise.

  Therefore, as a simpler method, Non-Patent Document 2 proposes a method of estimating a symbol on which extraneous noise is superimposed and processing the symbol with a bit likelihood of 0 during demodulation and decoding. In this method, when the bit likelihood input to the decoder is calculated, if the bit likelihood is greater than a predetermined threshold, it is determined that the bit is affected by a large amount of external noise, The bit likelihood is set to 0 and decoding is performed. By doing so, it is possible to perform reception in an external noise environment by a relatively simple method.

JP-A-1-190143 JP 2000-244464 A Japanese Patent No. 435632

Hideki Kanemoto, Shinichi Miyamoto, Norihiko Morinaga "Parameter estimation in class A impulse radio noise environment and error rate characteristics of optimal receiver", IEICE Transactions, Vol. J82-B, no. 12, pp. 2364-2374, December 1999 Liang Zhang and Abbas Yongacoglu, “Turbo Decoding with Erasure for High-Speed Transition in Impulse of Impulse, In 2002. 20-1-20-6, Feb. 2002. Changyan Di, David Proietti, I.D. Emre Telatar, Thomas J. et al. Richardson, and Rudiger L. Urbane, “Finite-Length Analysis of Low-Density Parity-Check Codes on the Binary Erasure Channel”, IEEE Transaction on Information. 48, no. 6, June 2002.

  By the way, in IEEE802.11n, which is expected to spread widely as a wireless LAN standard, IEEE802.16e of outdoor wireless broadband service, and some standards of power line communication systems, LDPC (Low-Density Parity-Check) is used as an error correction code. ) Code is adopted. The LDPC code is an error correction code defined by a low-density parity check matrix, and is known as an error correction code having a high error correction capability approaching the Shannon limit. Note that low density means that the number of non-zero elements in the parity check matrix is very small compared to the number of zero elements.

Sum-product decoding based on a reliability propagation algorithm is generally used for decoding an LDPC code. In the following, sum-product decoding in an AWGN channel with an average of 0 and a variance σ ² will be described.

FIG. 1 shows an encoding / decoding model in the AWGN communication channel. The K transmission information bits (b ₁ ,..., B _K ) are LDPC encoded by the LDPC encoder 1100 using the parity check matrix H, and N codeword bits (c ₁ ,. c _N ) is obtained. An example of the parity check matrix H is shown in Equation (1).

式（１）に示されるパリティ検査行列Ｈは、符号長１０、符号化率１／２のＬＤＰＣ符号を定義する５行１０列のパリティ検査行列である。但し、式（１）のパリティ検査行列Ｈは、表記し易くするために小さいサイズの行列としているので低密度な行列になっていない。

The parity check matrix H shown in Equation (1) is a parity check matrix of 5 rows and 10 columns that defines an LDPC code having a code length of 10 and a coding rate of 1/2. However, the parity check matrix H of the equation (1) is not a low-density matrix because it is a small-size matrix for easy description.

The encoding of the LDPC code using the parity check matrix H of Equation (1) is performed by an arbitrary algorithm that obtains a codeword C = (c ₁ ,..., C _N ) that satisfies HC ^T = 0, In the case of codes, (c ₁ ,..., C _N ) = (b ₁ ,..., B _K , p ₁ ,..., P _M ). Here, (p ₁ ,..., P _M ) are parity bits generated by LDPC encoding.

  FIG. 2 shows a Tanner graph representation of the parity check matrix H of Equation (1). The Tanner graph is composed of the same number of variable nodes as the number of columns of the parity check matrix H, the same number of check nodes as the number of rows of the parity check matrix H, and the same number of edges as the number of nonzero elements included in the parity check matrix H. 2 is a configured bipartite graph. The variable node and the check node are independent computing units. In the Tanner graph, when the element in the m-th row and the n-th column is non-zero in the parity check matrix H, the m-th check node and the n-th variable node are connected by an edge.

N codeword bits (c ₁ ,..., C _N ) are modulated by a modulator 1200. For BPSK (Binary Phase Shift Keying), modulation symbols _{x n} corresponding to the codeword bit _{c n is} given by x n = _-2c n +1. That is, when _{c n} is 0, _{x n} is 1, when _{c n} is 1, _{x n} is -1.

The modulation symbol is received by the receiving apparatus after being affected by the AWGN having an average of 0 and a variance σ ² in the AWGN communication path 1300. The received symbol y _n is given by y _n = x _n + z _n . z _n is an AWGN component.

Demodulator 1400, the received symbol _{y n,} log likelihood ratio of each codeword bit (LLR: Log Likelihood Ratio) obtaining the lambda _n. For BPSK, λ _n is given by λ _n = 2y _n / σ ² .

The LDPC decoder 1500 performs sum-product decoding with the LLR as an input. Let the parity check matrix H of M rows and N columns be the parity check matrix of the LDPC code to be decoded. However, K = N−M. Also, the element in the m-th row and the n-th column of the parity check matrix H is denoted as H _{m, n} . Subsets A (m) and B (n) of the set [1, N] are defined as shown in equations (2) and (3).

すなわち、部分集合Ａ（ｍ）はパリティ検査行列Ｈのｍ行目において値が１である列インデックスの集合を表し、部分集合Ｂ（ｎ）はパリティ検査行列Ｈのｎ列目において値が１である行インデックスの集合を表す。また、部分集合Ａ（ｍ）から要素ｎを除いた残りの要素ｎ’を式（４）と表す。

That is, the subset A (m) represents a set of column indexes whose value is 1 in the m-th row of the parity check matrix H, and the subset B (n) has a value of 1 in the n-th column of the parity check matrix H. Represents a set of row indexes. Further, the remaining element n ′ obtained by removing the element n from the subset A (m) is represented by Expression (4).

同様に、部分集合Ｂ（ｎ）から要素ｍを除いた残りの要素ｍ’を式（５）と表す。

Similarly, the remaining element m ′ obtained by removing the element m from the subset B (n) is represented by Expression (5).

ｓｕｍ−ｐｒｏｄｕｃｔ復号は以下のステップ１−６で実施される。

Sum-product decoding is performed in the following steps 1-6.

#Step 1 (Initialization)
The LDPC decoder 1500 sets a variable q, which is a counter of the number of iterations, to q = 1, and sets the maximum number of iterations to Q.

#Step 2 (column processing / processing at variable node)
In LDPC decoder 1500, each variable node, a LLRramuda _n inputted, decodes the code repeated by using the external value alpha _{m, n,} determine the priori value beta _{m, n.} Each variable node in advance uses the following update formula (6) for all pairs (m, n) that satisfy H _{m, n} = 1 in the order of n = 1, 2,. The value β _{m, n} is updated. However, only in the case of q = 1, each variable node performs the calculation of Expression (6) with α _{m, n} = 0, and when q ≠ 1, the external value α _{m, n} updated in step 3 described later. Equation (6) is calculated using the value of.

図３にタナーグラフ上での変数ノード処理の例を示す。図３において、ｎ＝５番目の変数ノードに接続されるのはｍ＝１、２、３番目の検査ノードである。これは、式（１）のパリティ検査行列Ｈの５列目において、Ｈ_ｍ，ｎ＝１を満たすのは、１、２、３行目であることに相当する。ｎ＝５番目の変数ノードは、ｍ＝１、２、３それぞれについて、式（６）を用いて事前値β_１，５、β_２，５、β_３，５を求める。事前値β_１，５、β_２，５、β_３，５は式（７）となる。但し、図３はｍ＝１の計算例を示す。

FIG. 3 shows an example of variable node processing on the Tanner graph. In FIG. 3, it is the m = 1, 2, and 3rd check nodes that are connected to the n = 5th variable node. This is equivalent to satisfying H _{m, n} = 1 on the first _, second, and third rows in the fifth column of the parity check matrix H of Equation (1). The n = 5th variable node obtains prior values β _1,5 , β _2,5 , β _3,5 for each of m = 1, ₂ , ₃ using equation (6). Prior values β _1,5 , β _2,5 , β _3,5 are given by equation (7). However, FIG. 3 shows a calculation example of m = 1.

ＬＬＲλ_ｎは、通信路から得られるｎ番目の符号語ビットが０であるか１であるかのビット尤度（確からしさ）であり、外部値α_ｍ，ｎは、ｍ番目の検査ノードから得られるｎ番目の符号語ビットが０であるか１であるかの尤度である。

LLRλ _n is the bit likelihood (probability) of whether the nth codeword bit obtained from the communication channel is 0 or 1, and the external value α _{m, n} is obtained from the mth check node. This is the likelihood that the nth codeword bit to be generated is 0 or 1.

The n-th variable node obtains the likelihood of whether the n-th codeword bit is 0 or 1 from LLRλ _n and α _{m ′, n,} and uses the result as a prior value β _{m, n} for the m-th variable node. Processing to send to the inspection node.

#Step 3 (Processing at the row processing / inspection node)
In LDPC decoder 1500, each check node decodes a single parity check code using a prior value β _{m, n} sent from a variable node to obtain an external value α _{m, n} . Each check node uses the following update formula (8) for all pairs (m, n) that satisfy H _{m, n} = 1 in the order of m = 1, 2,. The value α _{m, n} is updated.

図４にタナーグラフ上での検査ノード処理の例を示す。図４において、ｍ＝３番目の検
査ノードに接続されるのはｎ＝１、３、５、７、９、１０番目の変数ノードである。これは、式（１）のパリティ検査行列Ｈの３行目において、Ｈ_ｍ，ｎ＝１を満たすのは、１、３、５、７、９、１０列目であることに相当する。ｍ＝３番目の検査ノードは、ｎ＝１、３、５、７、９、１０それぞれについて、式（８）を用いて外部値α_３，１、α_３，３、α_３，５、α_３，７、α_３，９、α_３，１０を求める。但し、図４はｎ＝１の計算例を示す。

FIG. 4 shows an example of check node processing on the Tanner graph. In FIG. 4, n = 1, 3, 5, 7, 9, and 10th variable nodes are connected to the m = third check node. This is equivalent to satisfying H _{m, n} = 1 in the first _, third, fifth, seventh, ninth, and tenth columns in the third row of the parity check matrix H of Equation (1). The m = 3rd check node uses external values α _3,1 , α _3,3 , α _3,5 , α for each of n = ₁ , ₃ , ₅ , 7, 9, 10 using equation (8). _3,7 , _α3,9 , _α3,10 are obtained. However, FIG. 4 shows a calculation example where n = 1.

#Step 4 (Calculate temporary estimated words)
The LDPC decoder 1500 calculates the a posteriori LLRΛ _n for n∈ [1, N] by the equation (9), and then calculates the equation (10). In this way, the LDPC decoder 1500 calculates temporary estimated words (c ′ ₁ ,..., C ′ _N ).

＃ステップ５（パリティ検査）
ＬＤＰＣ復号器１５００は、一時推定語（ｃ’_１，・・・，ｃ’_Ｎ）が符号語になっているかどうかを検査する。つまり、ＬＤＰＣ復号器１５００は、一時推定語（ｃ’_１、・・・、ｃ’_Ｎ）が式（１１）を満たすかを判定する。

#Step 5 (parity check)
The LDPC decoder 1500 checks whether the temporary estimated word (c ′ ₁ ,..., C ′ _N ) is a code word. That is, LDPC decoder 1500 determines whether or not temporary estimated words (c ′ ₁ ,..., C ′ _N ) satisfy Expression (11).

一時推定語（ｃ’_１，・・・，ｃ’_Ｎ）が式（１１）を満たせば、一時推定語（ｃ’_１、・・・、ｃ’_Ｎ）は符号語になっているため、ＬＤＰＣ復号器１５００は一時推定語（ｃ’_１、・・・、ｃ’_Ｎ）を推定語として出力し、アルゴリズムを終了する。

If the temporary estimated words (c ′ ₁ ,..., C ′ _N ) satisfy Expression (11), the temporary estimated words (c ′ ₁ ,..., C ′ _N ) are codewords, The LDPC decoder 1500 outputs temporary estimated words (c ′ ₁ ,..., C ′ _N ) as estimated words, and ends the algorithm.

#Step 6 (Count iterations)
When the temporary estimated words (c ′ ₁ ,..., C ′ _N ) do not satisfy Expression (11), when q <Q, q is incremented and the process returns to Step 2. If q = Q, the LDPC decoder 1500 outputs temporary estimated words (c ′ ₁ ,..., c ′ _N ) as estimated words, and ends the algorithm.

The LDPC decoder 1500 outputs estimated words (c ′ ₁ ,..., C ′ _N ) obtained by performing the decoding algorithm in Step 1-6 described above. When systematic encoding is performed by the LDPC encoder 1100, the received information bits (b ′ ₁ ,..., B ′ _N ) are (b ′ ₁ ,..., B ′ _N ) = (c ′ ₁ ,..., C ′ _N ).

The main calculations of the sum-product decoding described above are an addition process at a variable node and an arithmetic process of tanh / tanh ^{−1 at} a check node. Since the cost for the tanh / tanh ⁻¹ arithmetic processing is higher than that for the addition processing, lowering the calculation cost for such arithmetic processing at the time of mounting leads to higher speed and compactness of the decoding method. As one method of speeding up and compacting, there is a method of approximating the tanh / tanh ⁻¹ arithmetic processing in sum-product decoding with a minimum value. This method is called min-sum decoding, and the check node uses equation (12) instead of equation (8).

また、変数ノードでは、式（６）の代わりに、式（１３）を用いる。

In the variable node, equation (13) is used instead of equation (6).

ここで、ｗは、１以下の正の実数である。（３、６）の正則ＬＤＰＣ符号の場合、ｗとして０．８程度の値を与えればｓｕｍ−ｐｒｏｄｕｃｔ復号と同等の誤り率特性が得られることが確認されている。但し、ｗの最適な値はパリティ検査行列Ｈによって異なる。

Here, w is a positive real number of 1 or less. In the case of the regular LDPC code of (3, 6), it has been confirmed that an error rate characteristic equivalent to that of sum-product decoding can be obtained if a value of about 0.8 is given as w. However, the optimum value of w varies depending on the parity check matrix H.

  Here, in order to remove the influence of the extraneous noise, it is considered that the bit likelihood of the bit expected to be affected by the extraneous noise as described above is set to 0 and input to the decoder. Hereinafter, the fact that the bit likelihood is 0 will be described as “erasure”. Similarly, setting the bit likelihood to 0 is expressed as “erasing”, “giving erasure”, and a bit having a bit likelihood of 0 as “erased bit”.

In FIG. 4, the case where n = 1, the bit corresponding to the fifth variable node is an erasure bit will be described as an example. The bit likelihood input to the variable node of the LDPC decoder 1500 is sent as a prior value β _{m, n} to a plurality of check nodes connected as they are at the start of decoding. Here, considering the processing in the check node 3 shown in FIG. 4, the check node 3 includes five prior values β _3,1 , β _3,3 , β _3,5 , β _3,7 , β _{3. , 9} and β _3,10 are input, of which the prior values β _3,1 and the prior values β _3,5 are zero. Since the check node operation represented by the equation (8) is multiplication of a value obtained by taking tanh of β _{m, n ′} input to the check node, two values are added to the prior value β _{m, n} input to the check node. If 0 is included, the resulting α _{m, n} is all 0.

tanh (0) = tanh ⁻¹ (0) = 0
Therefore, the external value is not updated in the calculation at the m = 3rd check node, and as a result, the iterative decoding gain decreases. This phenomenon occurs similarly even in the case of the equation (12) using the minimum value search instead of tanh / tanh ⁻¹ . Since the minimum value of the absolute value of β _{m, n ′} is returned as the external value α _{m, n , if} there are two or more 0s in β _{m, n} , the resulting α _{m, n} is all 0.

If the prior value β _3,1 and the prior value β _3,5 are updated to have a value other than 0 by processing of n = 1 and the fifth variable node in the subsequent iterative processing, the above-mentioned local The problem is solved, but the number of iterations required to obtain the desired bit error rate is increased, and the bit error rate that can be reached after a predetermined number of iterations is performed is deteriorated. Remains.

  Furthermore, considering the worst case, when erasure occurs in a pattern in which the external values output from all the check nodes are 0, the external values are not updated and the decoding performance is greatly deteriorated. . For example, this phenomenon occurs when all bits included in a set of codeword bits called a stopping set defined in Non-Patent Document 3 are lost. Also, when a plurality of bits that are part of the stopping set are lost, the decoding performance deteriorates if not all are lost.

  An example is shown using computer simulation results. FIG. 5 shows a simulation result (case (a)) of bit error rate characteristics after decoding when the second, fourth, and eighth bits are decoded as erasures in the parity check matrix H of equation (1), 6 is a graph showing a simulation result (case (b)) of bit error rate characteristics when the fourth, fifth, and seventh bits are decoded as erasures in the parity check matrix H of 1). Here, the combination of the 2nd, 4th, and 8th bits that are lost in the case (a) is a stopping set of the parity check matrix H, and the 4th, 5th, and 7th lost in the case (b). The bit combination is not a stopping set of the parity check matrix H. The decoding method used in the simulation is min-sum decoding with a maximum number of iterations of 30 and w = 0.8. In case (a), since the erasure bits constitute a stopping set, the bit error rate does not improve even if Eb / N0 is increased. On the other hand, in the case (b), since the erasure bits do not constitute a stopping set, the bit error rate is improved as Eb / N0 increases.

  As confirmed by the above simulation, in order to avoid the influence of the external noise in the external noise environment, the conventional configuration for decoding the bit affected by the external noise as the erasure includes the erasure bit. Even with the same number of bits, there is a problem that the decoding performance is significantly degraded depending on the combination of erasure bits.

  Further, Patent Document 1 and Patent Document 2 are conventional performance improvement means that consider error correction codes in an external noise environment.

  Patent Document 1 is to count the number of occurrences of impulsive noise before signal transmission and change the number of transmissions of transmission signals according to the number of occurrences. In this method, changing the number of transmissions corresponds to changing the number of repetitions (coding rate) of the repetition code. By using this method, the probability of successful transmission can be increased even when a large amount of impulsive noise is generated.

Further, Patent Document 2 describes the generation period of an impulsive noise so that the number of impulses of periodic impulsive noise included in one Reed-Solomon code block is within the number of correctable errors of the Reed-Solomon code. Controls the block length of the Reed-Solomon code. By using this method, it is possible to avoid occurrence of more errors than can be corrected within the block length of the Reed-Solomon code, and the error correction effect can be obtained efficiently.

  However, these two methods cannot solve the problem of deterioration in decoding performance due to the combination of erasures given in bit units.

  Accordingly, an object of the present invention is to provide a transmission device and a reception device, and a transmission method and a reception method that can suppress degradation in decoding performance due to a combination of erasures given in bit units.

  In order to solve the above-described problem, the transmission apparatus of the present invention has the same code length and the same coding rate, and the occurrence of external noise among a plurality of LDPC coding schemes defined by different parity check matrices. And a codeword bit sequence by encoding transmission data using the LDPC encoding method determined by the determination unit, and a determination unit that determines the LDPC encoding method to be used as the LDPC encoding method to be used. And an encoding unit for generating.

  According to this, even if the decoding performance of the receiving apparatus is deteriorated in the LDPC encoding scheme in which there is a combination of codeword bits that are lost due to the influence of extraneous noise, the LDPC code used for encoding transmission data Since the encoding method can be changed to another LDPC encoding method, it is possible to improve the decoding performance in the receiving apparatus without changing the code length and the encoding rate.

The figure which shows the model of the encoding / decoding in an AWGN communication channel. The figure which shows the Tanner graph corresponding to the parity check matrix H. The figure which shows the example of the variable node process on a Tanner graph. The figure which shows the example of the test | inspection node process on a Tanner graph. The figure which shows an example of the bit error rate characteristic in case there exists an erasure | elimination bit. The figure which shows an example of the system configuration | structure of the communication system 1 in Embodiment 1 of this invention. The figure which shows the frequency spectrum of the signal transmitted from the communication apparatus 2 of FIG. The figure which shows an example of the noise spectrum which the noise source 4 of FIG. 6 has generate | occur | produced. The figure which shows the structure of the communication apparatus 2 of FIG. The figure which shows the internal structure of the receiving characteristic estimation part 14 of FIG. The figure which shows the structure of the communication apparatus 3 of FIG. The figure which shows the structure of 2 A of communication apparatuses which have a function which changes the interleave pattern of the codeword bit series in the modification of Embodiment 1 of this invention. The figure which shows the structure of 3 A of communication apparatuses which perform power line communication with 2 A of communication apparatuses of FIG. The figure which shows the structure of the communication system 1B in the modification of Embodiment 1 of this invention. The figure which shows the structure of the communication apparatus 2B of FIG. The figure which shows the structure of the communication apparatus 3B of FIG. The figure which shows the structure of communication apparatus 2D in Embodiment 2 of this invention. FIG. 18 is a diagram showing an internal configuration of a reception characteristic estimation unit 14D in FIG. The figure which shows the noise level estimated in the noise generation condition estimation part 20 of FIG.

  A first transmission device which is one embodiment of the present invention has the same code length and the same coding rate, and has an external noise generation state among a plurality of LDPC coding methods defined by different parity check matrices. A determination unit that determines a corresponding LDPC encoding method as an LDPC encoding method to be used, and a codeword bit sequence by encoding transmission data using the LDPC encoding method determined by the determination unit An encoding unit to be generated.

  The first receiving apparatus according to one embodiment of the present invention has the same code length and the same coding rate, and the occurrence of external noise among a plurality of LDPC coding schemes defined by different parity check matrices A receiver that receives a signal that transmits a codeword bit sequence generated by encoding transmission data using an LDPC encoding method corresponding to the above, and demodulating the signal received by the receiver, A demodulator that generates a received codeword bit sequence corresponding to the codeword bit sequence; and a decoding process corresponding to the LDPC encoding method used for encoding the transmission data for the received codeword bit sequence And a decoding unit for performing.

  The first transmission method according to one aspect of the present invention is a method for generating extraneous noise among a plurality of LDPC coding methods having the same code length and the same coding rate and defined by different parity check matrices. A determination step of determining a corresponding LDPC encoding method as an LDPC encoding method to be used, and encoding of transmission data using the LDPC encoding method determined in the determination step An encoding step to be generated.

  The first reception method which is one embodiment of the present invention is the occurrence of extraneous noise among a plurality of LDPC coding schemes having the same code length and the same coding rate and defined by different parity check matrices. A reception step of receiving a signal transmitting a codeword bit sequence generated by encoding transmission data using an LDPC encoding method corresponding to the above, and demodulating the signal received in the reception step, A demodulation step for generating a received codeword bit sequence corresponding to the codeword bit sequence; and a decoding process corresponding to the LDPC encoding method used for encoding the transmission data for the received codeword bit sequence Performing a decoding step.

  According to the first transmission device, the first reception device, the first transmission method, and the first reception method, in the LDPC encoding scheme in which there is a combination of codeword bits that are lost due to the influence of external noise Even if the decoding performance of the receiving apparatus is deteriorated, the LDPC encoding method used for encoding transmission data can be changed to another LDPC encoding method, so the code length and the encoding rate must be changed. Therefore, the decoding performance in the receiving apparatus can be improved.

  A second transmission device according to an aspect of the present invention is the first transmission device, wherein the LDPC encoding scheme includes at least a first LDPC encoding scheme defined by a first parity check matrix, A second LDPC encoding scheme defined by a parity check matrix of 2 and constitutes a stopping set for the first parity check matrix, but for the second parity check matrix There are a predetermined number of codeword bit combinations that do not constitute a stopping set.

  According to the second transmission apparatus, even when the combination of codeword bits that are lost due to the influence of extraneous noise constitutes a stopping set for the first parity check matrix, the combination is a skip. By using the second LDPC encoding method defined by the second parity check matrix that does not constitute the topping set, the decoding performance in the receiving apparatus can be improved.

A third transmission device according to one embodiment of the present invention includes a second transmission device that includes codeword bits included in a codeword bit sequence generated by encoding based on an external noise generation state. A erasure pattern estimator that estimates the number of erasure candidate bits, which are codeword bits that are highly likely to be erased before the decoding process by the receiver, and the erasure pattern indicating the position of each of the erasure candidate bits in the codeword; The determination unit estimates a decoding performance in the parity check matrix from a relationship between the erasure pattern estimated by the erasure pattern estimation unit and the parity check matrix of the LDPC coding method, and based on an estimation result, The LDPC encoding method to be used is determined from a plurality of LDPC encoding methods.

  According to the third transmission apparatus, transmission data can be encoded by an LDPC encoding method defined by a parity check matrix having good decoding performance with respect to an erasure pattern estimated based on the occurrence of external noise. It becomes possible.

  According to a fourth transmission apparatus which is an aspect of the present invention, in the third transmission apparatus, the determination unit includes a combination of positions in each codeword of erasure candidate bits indicated by the erasure pattern as a stopping set. The LDPC encoding method defined by the parity check matrix not configured is determined as the LDPC encoding method to be used.

  According to the fourth transmission apparatus, transmission data is transmitted by an LDPC encoding method defined by a parity check matrix that does not constitute a stopping set with good decoding performance for a combination of positions in each codeword of the erasure candidate bit pattern. Can be encoded.

  A fifth transmission apparatus according to an aspect of the present invention is the fourth transmission apparatus, wherein the combination of positions in each codeword of the erasure candidate bit pattern indicated by the erasure pattern does not form a stopping set in the fourth transmission apparatus. A matrix means that at least one row having a row weight of 1 exists in a submatrix obtained by extracting a column corresponding to a position in each codeword of erasure candidate bits indicated by the erasure pattern from the parity check matrix. It is.

  According to a sixth transmitting apparatus which is an aspect of the present invention, in the third transmitting apparatus, the determining unit determines that a combination of positions of the erasure candidate bits indicated by the erasure pattern in each codeword has a stopping set. Of the parity check matrix that is not configured, the number of rows in which the row weight of the submatrix extracted from the column corresponding to the position in each codeword of the erasure candidate bits indicated by the erasure pattern is 0 or 1 is larger The LDPC encoding method defined by the parity check matrix is determined as the LDPC encoding method to be used.

  According to the sixth transmitting apparatus, a parity check matrix having a better decoding performance among parity check matrices that do not constitute a stopping set having a good decoding performance for a combination of positions in each codeword of the erasure candidate bit pattern. Transmission data can be encoded by the defined LDPC encoding method.

  In a seventh transmission device according to one aspect of the present invention, in the third transmission device, the determination unit determines that a combination of positions in each codeword of erasure candidate bits indicated by the erasure pattern has a stopping set. Of the parity check matrix not configured, the maximum value of the row weight of the submatrix extracted from the column corresponding to the position in each codeword of the erasure candidate bits indicated by the erasure pattern is defined by a smaller parity check matrix. The LDPC encoding method to be used is determined as the LDPC encoding method to be used.

  According to the seventh transmitting apparatus, a parity check matrix having a better decoding performance among parity check matrices that do not constitute a stopping set having a good decoding performance for a combination of positions in each codeword of the erasure candidate bit pattern. Transmission data can be encoded by the defined LDPC encoding method.

  An eighth transmission device according to one aspect of the present invention further includes a noise generation state estimation unit that estimates an external noise generation state based on a reception signal received from a communication path in the third transmission device, The erasure pattern estimation unit estimates the erasure pattern based on the external noise generation state estimated by the noise generation state estimation unit.

  According to the eighth transmission apparatus, the LDPC encoding method that uses the LDPC encoding method corresponding to the generation condition of the external noise only by the transmission apparatus without obtaining the generation condition of the external noise from another apparatus is used for encoding transmission data. It becomes possible to determine the encoding method.

  According to a ninth transmitting apparatus which is an aspect of the present invention, in the first transmitting apparatus, among the codeword bits constituting the codeword bit sequence, codeword bits which have been made erasure bits before decoding processing in the receiving apparatus. A reception unit that receives a signal including an erasure pattern indicating a position in each codeword; and the determination unit includes the erasure pattern received by the reception unit and the parity check matrix of the LDPC encoding method. From this relationship, the decoding performance in this LDPC encoding scheme is estimated, and the LDPC encoding scheme to be used is determined from the plurality of LDPC encoding schemes based on the estimation result.

  According to the ninth transmission device, each of the erasure pattern indicating the position in each codeword of the erasure candidate bits that can be estimated from the occurrence state of the external noise observed by the transmission device and each of the erasure bits lost on the receiving side. Even in a communication environment where the erasure pattern indicating the position of the codeword is different, the LDPC encoding method used for encoding the transmission data can be changed to an LDPC encoding method with good decoding performance in the receiving apparatus.

  According to a tenth transmitting apparatus which is an aspect of the present invention, in the first transmitting apparatus, among the codeword bits constituting the codeword bit sequence, codeword bits which have been made erasure bits before the decoding process in the receiving apparatus. A receiver that receives a signal including an LDPC encoding scheme determined by the receiver from a relationship between an erasure pattern indicating a position in each codeword and a parity check matrix of the LDPC encoding scheme; Determines the LDPC encoding method to be used based on the received signal including the LDPC encoding method.

  According to the tenth transmission device, each of the erasure pattern indicating the position in each codeword of the erasure candidate bits that can be estimated from the occurrence of the external noise observed by the transmission device and each of the erasure bits lost on the receiving side. Even in a communication environment where the erasure pattern indicating the position of the codeword is different, the LDPC encoding method used for encoding the transmission data can be changed to an LDPC encoding method with good decoding performance in the receiving apparatus.

  An eleventh transmission apparatus according to an aspect of the present invention is the first transmission apparatus, wherein the LDPC encoding scheme includes at least a first LDPC encoding scheme defined by a first parity check matrix, A second LDPC encoding method defined by two parity check matrices, and the second parity check matrix is a parity check matrix obtained by performing column replacement of the first parity check matrix Is a matrix equal to

  A twelfth transmitting apparatus according to an aspect of the present invention is the first transmitting apparatus, wherein the LDPC encoding scheme includes at least a first LDPC encoding scheme defined by a first parity check matrix, And a second LDPC encoding method defined by two parity check matrices, and the second parity check matrix is a matrix independent of the first parity check matrix.

A thirteenth transmission device according to one aspect of the present invention includes a modulation unit that generates a modulation signal by modulating a codeword bit sequence generated by the encoding unit in the first transmission device, and the modulation signal And a transmission unit for adding information indicating the LDPC encoding method used for encoding the transmission data by the encoding unit.

  According to the thirteenth transmission apparatus, the reception apparatus can perform a decoding process based on the LDPC encoding method used for encoding transmission data in the transmission apparatus.

  The fourteenth transmission device according to one aspect of the present invention determines to determine, as a replacement rule to be used, a replacement rule corresponding to the occurrence of external noise among a plurality of replacement rules for codeword bits in a codeword A first codeword bit sequence by encoding transmission data using an LDPC encoding method, and the replacement rule determined by the determination unit for the first codeword bit sequence And a coding unit that generates a second codeword bit sequence by exchanging codeword bits within the codeword.

  A second reception device which is one embodiment of the present invention uses a codeword in a codeword for a first codeword bit sequence generated by encoding transmission data using an LDPC encoding method. A receiving unit that receives a signal that transmits a second codeword bit sequence in which codeword bits in a codeword are replaced with a replacement rule corresponding to the occurrence of external noise among a plurality of replacement rules of bits; Demodulating a signal received by the receiving unit to generate a second received codeword bit sequence corresponding to the second codeword bit sequence; and a second received codeword bit sequence On the other hand, a first received codeword bit sequence is generated by exchanging bits according to an exchange rule opposite to the exchange rule used for exchanging codeword bits, and the first received codeword bit system And a decoding unit that performs decoding processing corresponding to the LDPC coding scheme respect.

  In the second transmission method according to one aspect of the present invention, a determination is made to determine, as a replacement rule to be used, a replacement rule corresponding to the occurrence of external noise among a plurality of replacement rules for codeword bits in a codeword. A first codeword bit sequence generated by encoding transmission data using an LDPC encoding method, and the replacement rule determined in the determining step for the first codeword bit sequence And a coding step for generating a second codeword bit sequence by exchanging codeword bits in the codeword.

  A second reception method according to one aspect of the present invention is a codeword in a codeword for a first codeword bit sequence generated by encoding transmission data using an LDPC encoding method. A reception step of receiving a signal that transmits a second codeword bit sequence in which codeword bits in a codeword are replaced with a replacement rule corresponding to the occurrence of external noise among a plurality of bit replacement rules; Demodulating the signal received in the receiving step to generate a second received codeword bit sequence corresponding to the second codeword bit sequence; and a second received codeword bit sequence On the other hand, a first received codeword bit sequence is generated by exchanging bits according to an exchange rule opposite to the exchange rule used for exchanging codeword bits, and the first received code Comprising a decoding step of performing decoding processing corresponding to the LDPC coding scheme with respect to a word bit sequence, the.

  According to the fourteenth transmission device, the second reception device, the second transmission method, and the second reception method, a code in a codeword having a combination of codeword bits that is lost due to the influence of extraneous noise Even if the word bit replacement rule deteriorates the decoding performance of the receiving device, the replacement rule to be used can be changed to another replacement rule, so that the decoding at the receiving device without changing the LDPC encoding method. The performance can be improved.

A fifteenth transmitting apparatus according to an aspect of the present invention includes: a determination unit that determines an interleave pattern corresponding to an external noise generation state among a plurality of interleave patterns for a codeword bit sequence as an interleave pattern to be used; An encoding unit that generates a codeword bit sequence by encoding transmission data using an encoding method, and the determination unit determines the codeword bit sequence generated by the encoding unit; An interleaving unit that generates a transmission bit sequence by performing interleaving with the interleaving pattern.

  The third receiving apparatus which is one embodiment of the present invention uses a codeword bit sequence generated by encoding transmission data using an LDPC encoding method to generate an exogenous noise among a plurality of interleave patterns. A receiver that receives a signal that transmits a transmission bit sequence interleaved with an interleave pattern corresponding to the demodulator, and a demodulator that generates a reception bit sequence corresponding to the transmission bit sequence by demodulating the signal received by the reception unit A deinterleaving unit that generates a received codeword bit sequence by performing deinterleaving with a data interleaving pattern opposite to the interleave pattern used for interleaving the codeword bit sequence with respect to the received bit sequence, A decoding process corresponding to the LDPC encoding method for the received codeword bit sequence. And a decoding unit that performs.

  A third transmission method according to an aspect of the present invention includes a determination step of determining an interleave pattern to be used as an interleave pattern to be used from among a plurality of interleave patterns for a codeword bit sequence, corresponding to the occurrence of external noise, and LDPC An encoding step for generating a codeword bit sequence by encoding transmission data using an encoding method, and a determination step for the codeword bit sequence generated in the encoding step. And an interleaving step of generating a transmission bit sequence by performing interleaving with the interleaving pattern.

The third reception method according to one aspect of the present invention is based on a codeword bit sequence generated by encoding transmission data using an LDPC encoding method, and the occurrence of extraneous noise among a plurality of interleave patterns. A reception step of receiving a signal that transmits a transmission bit sequence interleaved with an interleaving pattern corresponding to, and a demodulation that generates a reception bit sequence corresponding to the transmission bit sequence by demodulating the signal received in the reception step A deinterleaving step for generating a received codeword bit sequence by deinterleaving the received bit sequence with a data interleaving pattern opposite to the interleave pattern used for interleaving the codeword bit sequence; The LDPC encoding method for the received codeword bit sequence And a decoding step of performing decoding processing corresponding to.

  According to the fifteenth transmission device, the third reception device, the third transmission method, and the third reception method, the reception device in an interleave pattern in which there is a combination of codeword bits that is lost due to the influence of extraneous noise. Even if the decoding performance is degraded, the interleaving pattern to be used can be changed to another interleaving pattern, so that the decoding performance in the receiving apparatus can be improved without changing the LDPC encoding scheme. .

A sixteenth transmission device according to one aspect of the present invention provides codewords constituting a codeword bit sequence generated by encoding based on the occurrence of external noise among a plurality of subcarriers used for communication. Estimate subcarriers that transmit erasure candidate bits, which are codeword bits that are likely to be erased by the receiving device before decoding processing, and use a modulation scheme that is used for at least some of the estimated subcarriers. A determination unit that determines a tone map that uses a tone map that has a small multi-value modulation method among a plurality of modulation methods, and a codeword bit sequence by encoding transmission data using an LDPC encoding method An encoding unit to be generated, and modulating the codeword bit sequence generated by the encoding unit based on the tone map determined by the determination unit. And a modulator for generating a modulated signal.

  According to a fourth transmission method of one aspect of the present invention, among the plurality of subcarriers used for communication, each codeword constituting a codeword bit sequence generated by encoding based on the occurrence of external noise Estimate subcarriers that transmit erasure candidate bits, which are codeword bits that are likely to be erased by the receiving device before decoding processing, and use a modulation scheme that is used for at least some of the estimated subcarriers. A decision step for determining a tone map that uses a tone map having a modulation level with a small multi-level number among a plurality of modulation schemes, and encoding code of transmission data using an LDPC encoding scheme to convert a codeword bit sequence An encoding step to generate, and the codeword bit system generated in the encoding step based on the tone map determined in the determination step And a modulation step of generating a modulated signal by modulating.

  According to the sixteenth transmission device and the fourth transmission method, the number of codeword bits that are lost due to the influence of extraneous noise can be reduced, so that the reception device can change without changing the LDPC encoding method. The decoding performance can be improved.

  Embodiments of the present invention will be described below with reference to the drawings.

<< Embodiment 1 >>
Hereinafter, Embodiment 1 will be described with reference to the drawings.

  The communication device on the transmission side in the first embodiment can observe the occurrence of external noise occurring on the communication path, and can eliminate the observed occurrence of external noise before decoding by the communication device on the reception side Estimate erasure candidate bits that are highly reliable bits. Thereafter, the communication apparatus on the transmission side in Embodiment 1 receives the transmission data when the transmission data is encoded using the parity check matrix from the relationship between the estimated position of the erasure candidate bit in the code word and the parity check matrix. Parity check that estimates the decoding performance (hereinafter referred to as “decoding performance when using a parity check matrix” as appropriate) obtained by the communication apparatus of FIG. 1 and obtains a predetermined decoding performance based on the estimation result A matrix is obtained, and LDPC encoding of transmission data is performed using the obtained parity check matrix. However, the communication apparatus on the transmission side obtains a parity check matrix used for LDPC encoding of transmission data from a plurality of parity check matrices having the same number of rows and the same number of columns. This is because a parity check matrix used for encoding transmission data from a plurality of LDPC encoding schemes having the same code length and the same coding rate and defined by different parity check matrices. Is to obtain the LDPC encoding method defined by

  Note that here, the erasure candidate bits are expressed as bits that are highly likely to be lost in the communication device on the receiving side, except for the bits other than the bits determined as erasure candidate bits by the communication device on the transmission side, For example, there is a possibility of erasure in the communication device on the receiving side due to the influence of external noise that did not occur during the period when the occurrence of external noise was observed, and the bits determined to be erasure candidate bits are also in the reception state This is because there is a possibility that it may not be lost in the communication apparatus on the receiving side. The criteria for determining how much intensity and frequency the bit affected by the extraneous noise is determined as the erasure candidate bit based on the occurrence of the extraneous noise occurring in the communication path is the embodiment. 1 may be set in advance according to the type of extraneous noise assumed in the communication system according to one embodiment, other embodiments, and the communication system according to the modification, the number of assumed erasure candidate bits, and the like. The determination criterion may be selected according to the reception status.

FIG. 6 is a diagram illustrating an example of a system configuration of the communication system according to the first embodiment. The communication system 1 in FIG. 6 is a power line communication system in which a plurality of communication devices are connected to the same power line network, and power line communication is performed between the plurality of communication devices.

  The communication system 1 includes a communication device 2, a communication device 3, and a noise source 4, and the communication device 2 and the communication device 3 are each connected to a power line 5. Power line communication is performed between the communication device 2 and the communication device 3.

  The communication device 2 observes the occurrence of external noise generated on the communication channel, and the loss that is a bit that is highly likely to be lost before decoding in the receiving communication device from the observed occurrence of the external noise Candidate bits are estimated. And the communication apparatus 2 estimates the decoding performance when this parity check matrix is used from the relationship between the position in the codeword of the estimated erasure candidate bit and the parity check matrix while changing the parity check matrix, and estimates A parity check matrix capable of obtaining a predetermined decoding performance is obtained based on the result, and LDPC encoding of transmission data is performed using the obtained parity check matrix. However, the communication device 2 obtains a parity check matrix used for LDPC encoding of transmission data from a plurality of parity check matrices having the same number of rows and the same number of columns.

  Note that the communication apparatus 2 uses an LDPC encoding method corresponding to the occurrence of external noise from a plurality of LDPC encoding methods having the same code length and the same coding rate and defined by different parity check matrices. This is a form of a transmission apparatus that determines and encodes transmission data using the determined LDPC encoding method to generate an encoded bit sequence.

  When the communication device 3 receives a signal in an external noise environment, the communication device 2 erases bits estimated to be affected by the external noise, and then the communication device 2 performs LDPC encoding of transmission data. An LDPC decoding process based on the parity check matrix used is performed.

  The noise source 4 is an electric device that is a source of external noise connected to the same power line 5 as the communication device 2 and the communication device 3.

  FIG. 7 is a diagram illustrating a frequency spectrum of a signal transmitted from the communication device 2 of FIG. The communication device 2 uses a multicarrier modulation signal having a total bandwidth fw (Hz) and a subcarrier interval fsc (Hz) as a transmission signal. Examples of multicarrier modulation signals include OFDM (Orthogonal Frequency Division Multiplexing) using inverse discrete Fourier transform for multicarrier modulation and Wavelet-OFDM using inverse discrete wavelet transform, but for any type of multicarrier modulation. Applicable. The same applies to other embodiments and modifications using a multicarrier modulation signal.

  FIG. 8 is a diagram illustrating an example of a noise spectrum generated by the noise source 4 of FIG. In FIG. 8, only the noise power level at the center frequency of each subcarrier of the multicarrier modulation signal shown in FIG. 7 is shown.

  FIG. 9 is a diagram illustrating a configuration of the communication device 2 of FIG. The communication device 2 includes a coupling circuit 10, a reception unit 11, a demodulation unit 12, a decoding unit 13, a reception characteristic estimation unit 14, a transmission method determination unit 15, an encoding unit 16, a modulation unit 17, and a transmission unit 18.

  The coupling circuit 10 extracts a transmission signal superimposed on a power line that is a communication path and sends it to the reception unit 11. In addition, the coupling circuit 10 superimposes the transmission signal transmitted from the transmission unit 18 on a power line that is a communication path.

The reception unit 11 performs reception processing such as carrier detection, filtering processing, frequency conversion, synchronization, and A / D conversion of a signal (transmission signal extracted by the coupling circuit 10) transmitted through a power line that is a communication path. .

  When there is no carrier of a signal transmitted by another communication device on the power line, the coupling circuit 10 and the receiving unit 11 perform processing on noise on the power line, and the receiver 11 transmits to the demodulator 12. A signal related to noise will be sent.

  The demodulation unit 12 performs multicarrier demodulation processing on the reception signal sent from the reception unit 11 and designates the modulation method of each subcarrier of the multicarrier modulation signal used in the communication device that is the transmission source of the reception signal. The subcarrier is demodulated according to the tone map to calculate the bit likelihood of each transmitted codeword bit. Here, as described above, the demodulating unit 12 erases the code word bit determined to be strongly affected by the external noise with the bit likelihood being 0. The demodulator 12 sends the bit likelihood sequence to the decoder 13. For example, if the bit likelihood is greater than a predetermined threshold, the demodulator 12 determines that the bit is affected by extraneous noise having a large power. Further, the demodulation unit 12 performs multicarrier demodulation processing on the output signal of the reception unit 11 when there is no carrier of the signal transmitted by another communication device, and converts the signal obtained from the multicarrier demodulation result to noise The signal is sent to the reception characteristic estimation unit 14 as a situation estimation signal.

  The decoding unit 13 performs LDPC decoding processing based on the code length, coding rate, and parity check matrix of the LDPC code used in the transmission-side communication device, using the bit likelihood sequence sent from the demodulation unit 12. The resulting bit sequence is output as received data.

  Based on the noise situation estimation signal sent from the demodulator 12, the reception characteristic estimator 14 estimates erasure candidate bits that are highly likely to be lost before decoding in the communication device on the receiving side. When the parity check matrix is used from the relationship between the position of the estimated erasure candidate bit in the code word and the parity check matrix while changing the parity check matrix by performing column replacement, the reception characteristic estimation unit 14 The parity check matrix is obtained so that a predetermined decoding performance can be obtained based on the estimation result. However, the reception characteristic estimator 14 encodes all LDPC codes that may be specified by a control signal input to the transmission method determiner 15 (to be described later), coding rates thereof, and interleave patterns of codeword bit sequences. Then, a process of determining a parity check matrix so as to obtain a predetermined decoding performance is performed on the tone map.

  The reception characteristic estimator 14 performs a predetermined decoding on the code length of all LDPC codes that may be specified by the control signal, the coding rate, the interleave pattern of the codeword bit sequence, and the tone map. Instead of “determining the parity check matrix so that performance can be obtained”, instead of “the size of the minimum stopping set (minimum value of the number of bits constituting the stopping set), the average value of the stopping set size, the number The parity check matrix is determined so that a predetermined decoding performance can be obtained with respect to the code length and coding rate of the LDPC code in which the size of the most stopping set is smaller than the estimated number of erasure candidate bits. May be. In this case, the number of processes for determining the parity check matrix can be reduced.

  The reception characteristic estimation unit 22 recommends a transmission method including the obtained parity check matrix, the code length of the LDPC code used for obtaining the parity check matrix, the coding rate, the interleave pattern of the codeword bit sequence, and the tone map. The transmission method is sent to the transmission method determination unit 15.

  In the example shown in FIG. 9, the interleave pattern of the codeword bit sequence is only non-interleaved. However, when an interleave pattern other than the non-interleave pattern is to be included as the interleave pattern of the codeword bit sequence, the encoding unit 16 and the modulation unit 17 An interleaving unit that performs interleaving processing based on an interleaving pattern with respect to the encoded bit string may be provided.

  Note that since the parity check matrix is changed by column replacement, the parity check matrices to be estimated for decoding performance have the same number of rows and the same number of columns.

  Note that “changing the parity check matrix by performing column replacement” means that “in the case of a plurality of LDPC encoding methods having the same code length and the same coding rate and defined by different parity check matrices, It is one form of "changing an encoding system."

  The transmission method determination unit 15 is sent from the reception characteristic estimation unit 14 for all LDPC codes that may be specified by the control signal, their coding rates, interleave patterns of codeword bit sequences, and tone maps. The transmission data based on the recommended transmission method including the parity check matrix to be received and the instruction content in the input control signal (code length of LDPC code, coding rate thereof, interleave pattern of codeword bit sequence, tone map) A transmission method including a parity check matrix and the like actually used in the transmission is determined. Then, the transmission scheme determination unit 15 instructs the encoding unit 16 to perform LDPC encoding processing of transmission data with the code length, coding rate, and parity check matrix of the LDPC code included in the determined transmission scheme. Further, the transmission scheme determination unit 15 instructs the modulation unit 17 to modulate each subcarrier according to the tone map included in the determined transmission scheme.

  The encoding unit 16 performs LDPC encoding processing (error correction encoding processing) on the transmission data using the code length, encoding rate, and parity check matrix of the LDPC code specified by the transmission method determination unit 15. Thus, an encoded bit sequence is generated and output to the modulation unit 17.

  The modulation unit 17 maps the codeword bits transmitted from the encoding unit 16 to the modulation symbols by the modulation method of each subcarrier specified by the tone map specified by the transmission method determination unit 15 (for each subcarrier). Modulation). Further, the modulation unit 17 performs multicarrier modulation on a plurality of modulation symbols, generates a multicarrier modulation signal, and outputs the multicarrier modulation signal to the transmission unit 18.

  The transmission unit 18 adds a header or the like to the multicarrier modulation signal generated by the modulation unit 17, and further generates a transmission signal by performing transmission processing such as D / A conversion and frequency conversion, thereby coupling the coupling circuit 10. Send to.

  Note that the communication apparatus 2 can acquire information such as the code length, coding rate, and parity check matrix, interleave pattern, and tone map of the LDPC code used for actual transmission on the receiving side communication apparatus. The information is stored in the header and transmitted to the communication device on the receiving side. Instead of storing the information in the header and transmitting the information to the communication device on the reception side, the communication device 2 uses, for example, a signal different from the transmission signal in advance to transmit the information to the communication device on the reception side. You may make it transmit.

  FIG. 10 is a diagram showing an internal configuration of the reception characteristic estimation unit 14 of FIG. The reception characteristic estimation unit 14 includes a noise generation state estimation unit 20, a erasure pattern estimation unit 21, and a decoding performance estimation unit 22.

  The noise generation status estimation unit 20 estimates the noise generation status based on the noise status estimation signal sent from the demodulation unit 12. Noise generation conditions include the frequency (subcarrier number) at which strong external noise is generated, the frequency of generated external noise, the time interval of external noise generation, etc. A case where a subcarrier number in which extraneous noise occurs is estimated will be described as an example.

  The noise generation state estimation unit 20 detects a frequency band in which strong external noise is generated based on the noise state estimation signal, and subcarrier numbers fsc1, fsc2,..., Fscn corresponding to the frequency band. Ask for. Then, the noise generation state estimation unit 20 sends the obtained subcarrier numbers fsc1, fsc2,..., Fscn to the erasure pattern estimation unit 21.

  Based on the subcarrier numbers fsc1, fsc2,..., Fscn sent from the noise occurrence state estimation unit 20, the erasure pattern estimation unit 21 sets the bit likelihood to 0 during the decoding process of the receiving communication device. Number of erasure candidate bits (erasure candidate bits that are code word bits that are highly likely to be erased by the receiving communication apparatus before decoding) and positions (or position indices) of the erasure candidate bits. ). Hereinafter, an arrangement representing the number of erasure candidate bits and the position of each erasure candidate bit in the codeword is referred to as an erasure pattern. However, the erasure pattern estimation unit 21 determines the subcarrier number affected by extraneous noise, the tone map that specifies the modulation scheme of each subcarrier, the code length and coding rate of the LDPC code, and the codeword bit sequence. The disappearance pattern is estimated from the interleave pattern. The erasure pattern estimation unit 21 sends the estimated erasure pattern to the decoding performance estimation unit 22. Note that the erasure pattern estimation unit 21 includes code lengths of all LDPC codes that may be specified by the control signal input to the transmission method determination unit 15, coding rates thereof, interleave patterns of codeword bit sequences, and An erasure pattern is estimated for the tone map.

  The decoding performance estimation unit 22 changes the parity check matrix by performing column replacement, and uses the parity check matrix based on the relationship between the erasure pattern and the parity check matrix sent from the erasure pattern estimation unit 21. The decoding performance is estimated, and a parity check matrix that obtains a predetermined decoding performance based on the estimation result is obtained. However, the decoding performance estimation unit 22 obtains a parity check matrix that can obtain a predetermined decoding performance for all the erasure patterns sent from the erasure pattern estimation unit 21. Then, the reception characteristic estimation unit 14 includes the obtained parity check matrix, the code length of the LDPC code used for obtaining this, the coding rate thereof, the interleave pattern of the codeword bit sequence, and the tone map. The system is sent to the transmission system determination unit 15 as the recommended transmission system.

  Hereinafter, an example of the recommended transmission method determination operation performed by the communication device 2 will be described.

  The reception unit 11 performs reception processing on the signal transmitted from the coupling circuit 10 during non-communication, and then transmits the signal subjected to reception processing to the demodulation unit 12. The demodulation unit 12 performs multicarrier demodulation processing on the output signal of the reception unit 11 during non-communication. Since the output signal of the receiving unit 11 during non-communication does not include a transmission signal from another communication device, the demodulating unit 12 may acquire the noise level in each subcarrier from the result of the multicarrier demodulation process. it can. The demodulation unit 12 sends the noise level for each subcarrier acquired based on the result of the multicarrier demodulation process to the reception characteristic estimation unit 14 as a noise state estimation signal. Here, as the noise level, a value obtained from the output signal of the receiving unit 11 in a single multicarrier modulation signal symbol interval may be used, or the receiving unit 11 in a plurality of multicarrier modulation signal symbol intervals may be used. A statistical value such as a value obtained by averaging values obtained from the output signal or variance may be used.

  Further, during communication, noise is superimposed on the multicarrier modulation signal in the output signal of the receiving unit 11, so noise estimation accuracy is deteriorated as compared with non-communication, but the method described in Patent Document 3, for example, To obtain the noise level during communication.

The noise generation state estimation unit 20 in the reception characteristic estimation unit 14 estimates the subcarrier number in which strong external noise is generated based on the noise state estimation signal from the demodulation unit 12. Here, the case of the noise level shown in FIG. 8 will be described as an example. The noise generation situation estimation unit 20 compares the noise level in each subcarrier indicated by the noise situation estimation signal with the noise threshold Nth. The noise generation state estimation unit 20 derives the subcarrier numbers whose noise level is higher than the noise threshold Nth as fsc1, fsc2,. In the example of FIG. 8, if the subcarrier number is given as 1, 2,..., 10 in order from the lowest frequency in the total number of subcarriers 10, fsc1 = 2, fsc2 = 4, and fsc3 = 8. The noise generation state estimation unit 20 uses the information of the subcarrier number whose noise level is higher than the noise threshold Nth (in the example of FIG. 8, fsc1 = 2, fsc2 = 4, fsc = 8) as the noise generation state, and the erasure pattern estimation unit 21 Send to.

  The erasure pattern estimation unit 21 can cause the communication device on the reception side to erase before decoding from the noise generation status (subcarrier number information whose noise level is higher than the noise threshold Nth) sent from the noise generation status estimation unit 20. The number of erasure candidate bits that are high-quality codeword bits and the position of each erasure candidate bit in the codeword are estimated (an erasure pattern is estimated), and the estimated erasure pattern is sent to the decoding performance estimation unit 22. However, the erasure pattern estimation unit 21 determines the subcarrier number affected by extraneous noise, the tone map that specifies the modulation scheme of each subcarrier, the code length and coding rate of the LDPC code, and the codeword bit sequence. The disappearance pattern is estimated from the interleave pattern. Note that the erasure pattern estimation unit 21 includes code lengths of all LDPC codes that may be specified by the control signal input to the transmission method determination unit 15, coding rates thereof, interleave patterns of codeword bit sequences, and Estimate the erasure pattern for the tone map.

  In the following, description will be made under the condition that the multicarrier modulation signal uses BPSK for all subcarriers, the code length of the LDPC code is 10, and no interleaving is performed.

  A codeword obtained by LDPC encoding using an LDPC code having a code length of 10 is used for modulation of subcarriers from the lowest frequency in order from the codeword bit corresponding to the first column. That is, codeword bits corresponding to the first column are in subcarrier 1, codeword bits corresponding to the second column are in subcarrier 2,..., Codeword bits corresponding to the tenth column are in subcarrier 10. Assigned. Here, at the noise level shown in FIG. 8, since the subcarrier numbers where strong external noise is generated are fsc1 = 2, fsc2 = 4, and fsc3 = 8, the second, fourth, and eighth of the codewords. Are the erasure candidate bits that are highly likely to be erased by the receiving communication device before decoding. This is represented by an erasure pattern of the form shown in equation (14).

式（１４）に示す消失パターンにおいて、値が１となっている位置の符号語ビットは消失候補ビットであり、値が０となっている位置の符号語ビットは消失候補ビットではないことを表す。

In the erasure pattern shown in Expression (14), the code word bit at the position where the value is 1 is the erasure candidate bit, and the code word bit at the position where the value is 0 is not the erasure candidate bit. .

  The decoding performance estimation unit 22 calculates a value in the erasure pattern of the equation (14) sent from the erasure pattern estimation unit 21 from the parity check matrix H (code length 10, coding rate 1/2) of the equation (1). Only the columns (2 columns, 4 columns, 8 columns) corresponding to the codeword bits that are 1 are extracted, and a partial matrix Hp is created from the extracted columns. The created submatrix Hp is as shown in Equation (15).

式（１５）に示す部分行列Ｈｐは、行重み（各行の１の数）が１の行が１行もなく、行重みが０または２以上となっている。非特許文献３では作成される部分行列に行重みが１の行が存在しない場合に、その消失パターンはストッピングセットであると定義しており、式（１４）の消失パターンはストッピングセットを構成している。そのため、式（１４）の消失パターンとなった場合、図５で示したように誤り訂正復号後のビット誤り率は大幅に劣化する。復号性能推定部２２は、このように式（１４）の消失パターンに対して式（１）のパリティ検査行列Ｈを用いたときの復号性能が悪い（所定の復号性能が得られない）と判断する。復号性能推定部２２は、パリティ検査行列Ｈを用いたときの復号性能が悪いと判断すると、列置換を行うことによってパリティ検査行列を変更することによって復号性能の改善を図る。具体的には、消失ビットの組合せがストッピングセットを構成する場合に復号性能が劣化することが知られているため、復号性能推定部２２は、消失ビットの組み合わせがストッピングセットを構成しないように、パリティ検査行列Ｈの列置換を行う。

In the submatrix Hp shown in Expression (15), there is no row having a row weight (number of 1s in each row) of 1, and the row weight is 0 or 2 or more. In Non-Patent Document 3, when there is no row having a row weight of 1 in the created submatrix, the erasure pattern is defined as a stopping set, and the erasure pattern of Equation (14) is defined as a stopping set. It is composed. Therefore, when the erasure pattern of Expression (14) is obtained, as shown in FIG. 5, the bit error rate after error correction decoding is greatly deteriorated. In this way, the decoding performance estimation unit 22 determines that the decoding performance when the parity check matrix H of Expression (1) is used for the erasure pattern of Expression (14) is poor (predetermined decoding performance cannot be obtained). To do. If the decoding performance estimation unit 22 determines that the decoding performance is poor when the parity check matrix H is used, the decoding performance estimation unit 22 improves the decoding performance by changing the parity check matrix by performing column replacement. Specifically, since it is known that the decoding performance deteriorates when a combination of erasure bits constitutes a stopping set, the decoding performance estimation unit 22 does not make the combination of erasure bits constitute a stopping set. Next, column replacement of the parity check matrix H is performed.

  In the parity check matrix H expressed by the equation (1), the decoding performance estimation unit 22 performs the second column corresponding to the codeword bit that is 1 in the erasure pattern and the codeword bit that is 0 in the erasure pattern. The fifth column corresponding to the column is replaced, and the eighth column corresponding to the codeword bit which is 1 in the erasure pattern and the seventh column corresponding to the codeword bit which is 0 in the erasure pattern The parity check matrix H1 shown in Equation (16) is created by replacement.

復号性能推定部２２は、式（１６）のパリティ検査行列Ｈ１から、式（１４）の消失パターンにおいて１となっている符号語ビットに対応する列（２列、４列、８列）のみを抽出して、抽出した列から部分行列Ｈ１ｐを作成する。作成される部分行列Ｈ１ｐは式（１７）のようになる

The decoding performance estimation unit 22 extracts only columns (2 columns, 4 columns, 8 columns) corresponding to codeword bits that are 1 in the erasure pattern of Equation (14) from the parity check matrix H1 of Equation (16). Extraction is performed to create a submatrix H1p from the extracted columns. The created submatrix H1p is as shown in Equation (17).

式（１７）に示す部分行列Ｈ１ｐは、行重み１の行が２行あるため、パリティ検査行列Ｈ１に対して式（１４）の消失パターンが示す消失候補ビットの組み合わせはストッピングセットを構成していない。そのため、列置換したパリティ検査行列Ｈ１を用いることで、復号性能は改善される。復号性能推定部２２は、式（１４）の消失パターンに対して式（１６）のパリティ検査行列Ｈ１を用いたときの復号性能が良い（所定の復号性能が得られる）と判断する。

Since the submatrix H1p shown in the equation (17) has two rows of the row weight 1, the combination of the erasure candidate bits indicated by the erasure pattern of the equation (14) with respect to the parity check matrix H1 forms a stopping set. Not. Therefore, decoding performance is improved by using the parity check matrix H1 with column replacement. The decoding performance estimation unit 22 determines that the decoding performance is good (predetermined decoding performance is obtained) when the parity check matrix H1 of Expression (16) is used for the erasure pattern of Expression (14).

  The decoding performance estimation unit 22 performs the parity check matrix H1 subjected to column replacement, the code length 10 of the LDPC code used for obtaining this, the coding rate 1/2, and the codeword bit sequence interleaving, and A recommended transmission method including using BPSK in all subcarriers is sent to the transmission method determination unit 15. The decoding performance estimation unit 22 is configured to send the parity check matrix H1 itself to the transmission scheme determination unit 15, but is not limited thereto, and transmits an index indicating the parity check matrix H1. Alternatively, only the column replacement rule for the parity check matrix H (for example, information indicating a combination of columns to be replaced) may be sent.

  However, the decoding performance estimation unit 22 obtains a parity check matrix that obtains a predetermined decoding property for all the erasure patterns sent from the erasure pattern estimation unit 21, and sets the recommended transmission method to the transmission method determination unit 15. Send to.

  Here, the processing of the decoding performance estimation unit 22 is summarized.

#Step A1
The decoding performance estimation unit 22 converts the parity check matrix (first parity check matrix H, second and subsequent parity check matrices obtained by column replacement in step A3) into codeword bits that are 1 in the erasure pattern. Extract only the corresponding columns and create a submatrix from the extracted columns.

#Step A2
The decoding performance estimation unit 22 determines whether a combination of positions of the erasure candidate bits in each codeword constitutes a stopping set based on the partial matrix created in step A1.

#Step A3
When the combination of positions in each codeword of the erasure candidate bits does not form a stopping set, the decoding performance estimation unit 22 determines a recommended transmission method including the parity check matrix to be processed in Step A1. Send to part 15. On the other hand, the decoding performance estimation unit 22 performs column replacement of the parity check matrix H when the combination of positions of the erasure candidate bits in each codeword constitutes a stopping set, and performs the process of step A1. However, column replacement of the parity check matrix includes at least one column corresponding to a codeword bit which is 1 in the erasure pattern, at least one column corresponding to a codeword bit which is 0 in the erasure pattern, This is done by replacing Note that the parity check matrix to be replaced is always the parity check matrix H. For example, the first column replacement performs column replacement on the parity check matrix H, and the second and subsequent column replacements are performed in the previous column replacement. Column replacement may be performed on the parity check matrix obtained by the above.

  Depending on the erasure pattern, there may be no column replacement pattern that does not constitute a stopping set. In this case, for example, a default (no replacement) parity check matrix is used, or finally the stopping set is used. The parity check matrix for which it is determined whether or not to configure is used. Alternatively, a flag that the code length and the coding rate are not used is set, and the flag that is not used is subsequently made valid until the reception characteristic estimation unit 14 determines the next parity check matrix.

  In the first embodiment, the reception characteristic estimation unit 14 is configured so that the parity check matrix is changed by performing column replacement and the parity check matrix is used from the relationship between the erasure pattern and the parity check matrix. Although a configuration is employed in which the decoding performance is estimated and a parity check matrix that obtains a predetermined decoding performance is obtained, the configuration of the reception characteristic estimation unit 14 is not limited to this, and for example, is as follows. May be. For each of a plurality of parity check matrices having the same number of rows and the same number of columns and different matrix elements, the decoding performance in the parity check matrix for each of a plurality of erasure patterns is obtained in advance and received. The characteristic estimation unit 14 associates a parity check matrix having the best decoding performance among a plurality of parity check matrices with respect to each erasure pattern and stores the parity check matrix in a table. Then, the reception characteristic estimation unit 14 estimates the noise generation status based on the noise status estimation signal from the demodulation unit 12, estimates the erasure pattern from the estimated noise generation status, and estimates the erasure pattern with reference to the table In contrast, the parity check matrix having the best decoding performance may be read. By doing in this way, the communication apparatus 2 does not need to perform the process of creating a partial matrix according to the erasure pattern and estimating the decoding performance every time the noise occurrence state is observed. The processing load can be reduced.

  The communication device 2 performs the recommended parity check matrix determination process when not communicating or when necessary during communication. The necessary time during communication is, for example, when the frequency of error occurrence is high or when communication is interrupted.

  Hereinafter, an example of a transmission operation of transmission data performed by the communication apparatus 2 using the recommended transmission method sent from the reception characteristic estimation unit 14 to the transmission method determination unit 15 will be described.

  When the communication device 2 transmits transmission data, first, the transmission method determination unit 15 actually uses the input content from the decoding performance estimation unit 22 in the reception characteristic estimation unit 14 and the instruction content in the control signal. Determine a parity check matrix.

  For example, the control signal specifies that the code length of the LDPC code is 10, the coding rate is 1/2, the modulation scheme of all subcarriers is BPSK, and no interleaving is performed. Suppose that the parity check matrix included in the recommended transmission scheme from which is transmitted is the parity check matrix H1. In this case, the transmission scheme determination unit 15 instructs the encoding unit 16 to perform LDPC encoding processing of transmission data with the code length 10 of the LDPC code, the code rate 1/2, and the parity check matrix H1. In addition, the transmission scheme determination unit 15 instructs the modulation unit 17 to perform modulation of each subcarrier by BPSK.

The encoding unit 16 divides the transmission data into information block lengths based on the code length and coding rate of the LDPC code instructed from the transmission method determining unit 15, and determines the transmission method for the transmission data for each information block length. The codeword bit sequence is generated by performing the LDPC encoding process using the parity check matrix H1 instructed by the unit 15. Then, the encoding unit 16 sends the generated codeword bit sequence to the modulation unit 17.

  Since the modulation scheme of all the subcarriers instructed from the transmission scheme determination section 15 is BPSK, the modulation section 17 uses BPSK for all the subcarriers for the codeword bit sequence sent from the encoding section 16. And a multicarrier modulation signal is generated by performing multicarrier modulation on the result. Then, the modulation unit 17 sends the generated multicarrier modulation signal to the transmission unit 18.

  The transmission unit 18 performs transmission processing on the multicarrier modulation signal and then transmits the signal to the coupling circuit 10. The coupling circuit 10 superimposes the transmission signal on the power line and transmits it to the communication device 3.

  FIG. 11 is a diagram illustrating a configuration of the communication device 3 of FIG. The communication device 3 includes a coupling circuit 30, a receiving unit 31, a demodulating unit 32, and a decoding unit 33. In FIG. 11, only the configuration necessary for the description of the first embodiment is shown, but a configuration for transmitting a signal may actually be provided. Further, the configuration may be the same as that of the communication device 2 illustrated in FIG.

  The coupling circuit 30 extracts a transmission signal superimposed on a power line that is a communication path, and sends it to the reception unit 31.

  The reception unit 31 performs reception processing such as carrier detection, filtering processing, frequency conversion, synchronization, and A / D conversion of a signal (transmission signal extracted by the coupling circuit 30) transmitted through a power line that is a communication path. .

  The demodulation unit 32 performs multicarrier demodulation processing on the reception signal transmitted from the reception unit 31.

  The demodulator 32 performs a demodulation process for each subcarrier on the header portion. A header analysis unit (not shown) analyzes the processing result for the header part by the demodulation unit 32 and acquires information transmitted in the header part. For example, when information such as the code length of the LDPC code used for actual transmission, its coding rate, its parity check matrix, interleave pattern, and tone map is added to the header and transmitted, the header analyzer As a result of the analysis, the communication apparatus 3 can acquire the code length of the LDPC code used for actual transmission, its coding rate, its parity check matrix, interleave pattern, and tone map. In the configuration of the communication device 2 shown in FIG. 9 and the communication device 3 shown in FIG. 10, the interleave pattern is non-interleaved. However, when it is desired to include an interleave pattern other than the non-interleave pattern as the interleave pattern, the demodulator 32 and the decoder A de-interleaving unit that performs de-interleaving processing based on a de-interleaving pattern opposite to the interleaving pattern for the bit likelihood sequence may be provided.

  The demodulation unit 32 performs demodulation processing for each subcarrier on the payload portion according to the tone map used in the transmission-side communication apparatus, and calculates the bit likelihood of each received codeword bit. Here, as described above, the demodulating unit 32 erases the code word bit determined to be strongly influenced by the external noise with the bit likelihood being zero. The demodulator 32 sends the bit likelihood sequence to the decoder 33. For example, if the bit likelihood is larger than a predetermined threshold, the demodulator 32 determines that the bit is affected by external noise having a large power.

  The decoding unit 33 performs LDPC decoding processing based on the code length, coding rate, and parity check matrix of the LDPC code used in the transmission side communication device, using the bit likelihood sequence sent from the demodulation unit 32. The resulting bit sequence is output as received data. Here, the external noise affected by the communication device 3 has a correlation with the external noise affected by the communication device 2 on the principle of noise generation. Then, the communication device 2 considers the erasure pattern given by the demodulator 32 of the communication device 3 based on the external noise estimated by itself, and then determines the parity check matrix actually used for encoding the transmission data. The transmission signal is transmitted by encoding the transmission data using the determined parity check matrix. Therefore, even if the bit affected by the external noise is lost in the demodulation unit 32 (even if the bit likelihood is 0), the decoding performance in the decoding unit 33 is less deteriorated.

  According to the communication device 2 and the communication device 3 described above, even if the decoding performance of the reception device is deteriorated in a parity check matrix having a combination of codeword bits that is lost due to the influence of extraneous noise, Since the parity check matrix used for data encoding can be changed to another parity check matrix, the decoding performance in the receiving apparatus can be improved without changing the code length and coding rate.

  Hereinafter, supplementary contents and modifications of the first embodiment will be described.

  (1) An implementation method of LDPC encoding processing using a parity check matrix and LDPC decoding processing based on the parity check matrix will be described.

  (1-1) As a method of implementing encoding processing using a parity check matrix, the encoding unit 16 holds all the parity check matrices that may be used, and in accordance with an instruction from the transmission method determination unit 15 A configuration in which one to be used is selected from the held parity check matrix and transmission data is LDPC-encoded can be employed.

  Also, as a method of implementing decoding processing based on the parity check matrix, the decoding unit 33 holds all parity check matrices that may be used, and holds them according to the contents sent from the communication device on the transmission side. A configuration in which one of the parity check matrices to be used is selected and LDPC decoding processing is performed on the likelihood bit sequence can be employed.

  However, in this implementation method, the encoding unit 16 and the decoding unit 33 need to hold all parity check matrices that may be used, and the circuit scale of the encoding unit 16 and the decoding unit 33 increases. For example, you may employ | adopt two other mounting methods (1-2) and (1-3) described below.

  (1-2) The encoding unit holds only the basic parity check matrix H, and the transmission method determination unit 15 determines the column replacement rule for the parity check matrix H (from the parity check matrix H to the LDPC code of the actual transmission data). Column replacement rule to obtain a parity check matrix to be used for conversion. The encoding unit performs column replacement on the parity check matrix H according to the rule of column replacement for the parity check matrix H sent from the transmission method determination unit 15, and uses the parity check matrix subjected to column replacement to perform LDPC of transmission data. Encoding is performed.

  The decoding unit holds only the basic parity check matrix H, and the communication device on the transmission side uses the column replacement rule for the parity check matrix H (the parity check matrix used for encoding the actual transmission data from the parity check matrix H). Send column replacement rules to get). The decoding unit performs column replacement on the parity check matrix H according to the rule of column replacement for the parity check matrix sent from the communication device on the transmission side, and performs LDPC decoding on the likelihood bit sequence based on the parity check matrix subjected to column replacement Process.

  In this way, the encoding unit and the decoding unit need only hold the basic parity check matrix H, and therefore the circuit scale of the encoding unit and the decoding unit can be reduced.

  (1-3) The encoding unit does not perform LDPC encoding processing using a parity check matrix obtained by performing column replacement on the basic parity check matrix H, but uses LDPC using the basic parity check matrix H. The codeword bit sequence obtained by performing the encoding process, and the codeword within the codeword according to the replacement rule corresponding to the column replacement rule for the parity check matrix H sent from the transmission method determination unit 15 Bits may be exchanged. In this way, the LDPC encoding process performed by the encoding unit is only the LDPC encoding process using the basic parity check matrix H, and the circuit scale of the encoding unit can be reduced. For example, if the column replacement rule for the parity check matrix H is a rule that replaces the Kth column and the Lth (K ≠ L) column, the replacement rule corresponding to this column replacement rule is in the codeword. The rule is to replace the Kth codeword bit and the Lth codeword bit.

  However, this implementation method states that “a codeword bit sequence codeword using a codeword bit replacement rule in a codeword corresponding to a noise occurrence situation among codeword bit replacement rules in a plurality of codewords”. Code bit replacement rule is determined, transmission data is LDPC encoded to generate a coded bit sequence, and code word bits within the code word are determined according to the replacement rule determined for the generated coded bit sequence It is a form of “replacement”. The column replacement rule for the basic parity check matrix H corresponds to the codeword bit replacement rule in the codeword.

  Also, the decoding unit does not perform LDPC decoding processing based on a parity check matrix obtained by performing column replacement on the basic parity check matrix H, but performs basic parity check matrix H transmitted from the communication device on the transmission side. Based on the parity check matrix H using the bit likelihood sequence in which the bit is replaced by performing bit replacement on the likelihood bit sequence according to the bit replacement rule and the reverse replacement rule corresponding to the column replacement rule for It may be configured to perform LDPC decoding processing. In this way, the decoding process performed by the decoding unit is only the LDPC decoding process based on the basic parity check matrix H, and the circuit scale of the decoding unit can be reduced.

  However, in this modification example, “the bit likelihood sequence is replaced with the reverse replacement rule used for the replacement of the codeword bits, and the bit likelihood sequence is changed. The LDPC decoding process based on the parity check matrix is performed ”.

  In (1-2) and (1-3), only one basic parity check matrix is used. However, even if there are two or more basic parity check matrices, it can be easily expanded. .

  (2) In Embodiment 1, when a combination of positions in each codeword of erasure candidate bits indicated by an erasure pattern for a parity check matrix forms a stopping set, a parity check matrix is obtained by performing column replacement. In the above description, the combination is changed so that the combination does not constitute the stopping set. However, the present invention is not limited to this.

  For example, the parity check matrix is changed in order to improve decoding performance even when the combination of positions in each codeword of the erasure candidate bits indicated by the erasure pattern for the parity check matrix does not constitute a stopping set. Also good.

Specifically, the decoding performance is such that the row weight of the submatrix is 0 or 1, and the row weight is 2
Even in the above rows, there is a tendency that the smaller the row weight, the better. Therefore, based on this tendency, the row weight of the submatrix is 0 or 1, or the maximum value of the row weight of the submatrix is A method of performing column replacement of the parity check matrix so as to be smaller can be considered.

  An example for realizing this will be described. For each of the plurality of parity check matrices, the communication device 2 extracts a column corresponding to each of the erasure candidate bits indicated by the erasure pattern from the parity check matrix to create a partial matrix. Then, the communication device 2 is based on the partial matrix for each of the plurality of parity check matrices, and the combination of positions in each codeword of the erasure candidate bits indicated by the erasure pattern is a parity check matrix that does not form a stopping set. Among them, a parity check matrix having more rows with 0 or 1 row weights of the submatrix is determined to be used for LDPC encoding of transmission data. Alternatively, the communication device 2 can generate a parity check matrix in which a combination of positions in each codeword of erasure candidate bits indicated by the erasure pattern does not form a stopping set based on a partial matrix for each of a plurality of parity check matrices. Of these, a parity check matrix having a smaller maximum row weight of the submatrix is determined to be used for LDPC encoding of transmission data.

  (3) In Embodiment 1, the configuration for improving the decoding performance by column replacement of the parity check matrix has been described, but the configuration is not limited to this configuration. For example, when it is determined that the decoding performance for the estimated erasure pattern is poor, the codeword bit determined as the erasure candidate bit is replaced with another codeword bit by changing the interleave pattern of the codeword bit sequence Alternatively, the configuration may be such that improvement in decoding performance is performed. In this way, the encoding unit of the communication apparatus does not need to hold a plurality of parity check matrices, and similarly, the decoding unit of the communication apparatus does not need to hold a plurality of parity check matrices.

  Hereinafter, a communication device 2A having a function of changing an interleave pattern of a codeword bit sequence and a communication device 3A that performs power line communication with the communication device 2A will be described with reference to FIGS.

  Compared with the communication apparatus 2 described with reference to FIGS. 9 and 10, the communication apparatus 2 </ b> A having the function of changing the interleave pattern of the codeword bit sequence illustrated in FIG. 12 is different from the reception characteristic estimation section 14 in the decoding performance estimation section. In this configuration, instead of 22, a reception characteristic estimation unit 14A having a decoding performance estimation unit 22A is replaced, a transmission method determination unit 15 is replaced by a transmission method determination unit 15A, and a deinterleaving unit 51 and an interleaving unit 52 are added. However, since the interleave pattern of the codeword bit sequence is changed, for the sake of simplicity, the erasure when the erasure pattern estimation unit 21 in the reception characteristic estimation unit 14A does not perform the interleave will be described below. It is assumed that the pattern is estimated.

The decoding performance estimation unit 22A performs a deinterleaving process on the erasure pattern sent from the erasure pattern estimation unit 21 according to the deinterleave pattern. The decoding performance estimation unit 22A receives the interleaved codeword bit sequence according to the interleave pattern opposite to the deinterleave pattern from the relationship between the erasure pattern deinterleaved according to the deinterleave pattern and the parity check matrix. The decoding performance in the communication device is estimated. When the estimated decoding performance is poor (when a predetermined decoding performance cannot be obtained), the decoding performance estimation unit 22A changes the erasure pattern subjected to the deinterleaving process using a different deinterleaving pattern. In order to improve decoding performance. The decoding performance estimation unit 22A obtains a deinterleave pattern in which the decoding performance is improved (a predetermined decoding performance is obtained), and obtains an interleave pattern opposite to the obtained deinterleave pattern. However, the decoding performance estimation unit 22A obtains an interleave pattern that can obtain a predetermined decoding performance for all the erasure patterns sent from the erasure pattern estimation unit 21. Then, the reception characteristic estimator 14A determines the transmission method using the transmission method including the obtained interleave pattern, the code length of the LDPC code used for obtaining the same, the coding rate, and the tone map as the recommended transmission method. Send to part 15A.

  Below, the process of the decoding performance estimation part 22A is further demonstrated.

#Step B1
The decoding performance estimation unit 22A performs deinterleaving processing on the erasure pattern according to the deinterleaving pattern. As the deinterleave pattern, a different deinterleave pattern is used every time step B1 is performed.

#Step B2
The decoding performance estimation unit 22A extracts only the column corresponding to the codeword bit that is 1 in the erasure pattern deinterleaved in step B1 from the parity check matrix H, and creates a partial matrix from the extracted column.

#Step B3
Based on the partial matrix created in step B2, the decoding performance estimation unit 22A determines whether or not combinations of positions in each codeword of erasure candidate bits constitute a stopping set.

#Step B4
When the combination of positions of the erasure candidate bits in each codeword does not constitute a stopping set, the decoding performance estimation unit 22A interleaves the deinterleave pattern opposite to the deinterleave pattern used for the deinterleave of the erasure pattern in step B1. And a recommended transmission method including the obtained interleave pattern is sent to the transmission method determination unit 15A. On the other hand, the decoding performance estimation unit 22A performs the process of step B1 when the combination of positions in each codeword of the erasure candidate bits constitutes a stopping set.

  Depending on the erasure pattern, there may be no deinterleaving pattern that does not constitute a stopping set. In this case, for example, a default determinating (whether or not to configure a stopping set first) is performed. An interleave pattern opposite to the interleave pattern is used, or an interleave pattern opposite to the deinterleave pattern for which it is finally determined whether or not to configure a stopping set. Alternatively, a flag that the code length and the coding rate of the parity check matrix H are not used is set, and the flag that is not used is enabled until the reception characteristic estimation unit 14A next determines the next deinterleave pattern.

  When combinations of positions in the codewords of the erasure candidate bits indicated by the erasure pattern after deinterleaving with respect to the parity check matrix H form a stopping set, the combination can be changed by changing the deinterleaving pattern. The configuration in which the topping set is not configured has been described as an example, but is not limited thereto.

  For example, in order to improve the decoding performance even if the combination of the positions of the erasure candidate bits indicated by the erasure pattern after deinterleaving for the parity check matrix H does not form a stopping set, the erasure pattern data can be decoded. The interleave pattern (that is, the interleave pattern of the codeword bit sequence) may be changed.

  Specifically, the decoding performance tends to be better when the row weight of the submatrix is 0 or 1, and the row weight is 2 or more. The deinterleave pattern of the erasure pattern (that is, the interleave pattern of the codeword bit sequence) so that the number of row weights of the submatrix is 0 or 1 is larger or the maximum value of the row weight of the submatrix is smaller. A method of changing the value is conceivable.

  An example for realizing this will be described. For each of the plurality of deinterleave patterns, communication apparatus 2A extracts a column corresponding to each of the erasure candidate bits indicated by the erasure pattern after deinterleaving from parity check matrix H and creates a partial matrix. Then, communication device 2A, based on the partial matrix for each of the plurality of deinterleave patterns, does not form a stopping set in which the combination of the positions of the erasure candidate bits indicated by the erasure pattern after deinterleave in each codeword Of the deinterleave patterns, an interleave pattern opposite to the deinterleave pattern having more rows with 0 or 1 row weights in the submatrix is determined to be used for interleaving the codeword bit sequence. Alternatively, the communication device 2A determines the maximum value of the row weight of the submatrix among the deinterleave patterns in which the combination of the positions of the erasure candidate bits indicated by the erasure pattern after deinterleaving does not form a stopping set. It is decided to use an interleave pattern opposite to a deinterleave pattern smaller than for interleaving a codeword bit sequence.

  In the above description, the reception characteristic estimation unit 14A is configured to estimate the decoding performance from the relationship between the deinterleaved erasure pattern and the parity check matrix while changing the deinterleave pattern, and obtain a predetermined decoding performance. A configuration is employed in which an interleave pattern is obtained and an interleave pattern opposite to the obtained deinterleave pattern is obtained. However, the configuration of the reception characteristic estimation unit 14A is not limited to this, and for example, the following may be used. Good. For each of the plurality of erasure patterns, the decoding performance in the parity check matrix for the erasure pattern after deinterleaving based on each of the plurality of deinterleaving patterns is obtained in advance, and the reception characteristic estimation unit 14A performs for each erasure pattern A deinterleave pattern having the best decoding performance among a plurality of deinterleave patterns is stored in the table in correspondence. Then, the reception characteristic estimation unit 14A estimates the noise generation status based on the noise status estimation signal from the demodulation unit 12, estimates the erasure pattern from the estimated noise generation status, and estimates the erasure pattern estimated with reference to the table In contrast, a deinterleave pattern with the best decoding performance may be read, and an interleave pattern opposite to the read deinterleave pattern may be obtained. By doing in this way, the communication apparatus 2A does not need to perform the process of creating a partial matrix according to the erasure pattern and estimating the decoding performance every time the noise occurrence state is observed. The processing load can be reduced. The information corresponding to the erasure pattern may be, for example, an interleave pattern applied to the codeword bit sequence (a pattern opposite to the deinterleave pattern applied to the erasure pattern) instead of the deinterleave pattern applied to the erasure pattern. .

  The transmission method determination unit 15A determines the recommended transmission method corresponding to the instruction content in the control signal as the transmission method to be actually used. Then, the transmission scheme determination unit 15A instructs the encoding unit 16 to perform LDPC encoding processing of transmission data with the code length, coding rate, and parity check matrix of the LDPC code included in the determined transmission scheme. In addition, the transmission scheme determining unit 15A instructs the interleaving unit 52 to perform interleaving of the codeword bit sequence with the interleave pattern included in the determined transmission scheme. Furthermore, the transmission method determination unit 15A instructs the modulation unit 17 to modulate each subcarrier according to the tone map included in the transmission method.

  The interleaving unit 52 performs interleaving on the codeword bit sequence output from the encoding unit 16 in accordance with the interleaving pattern instructed from the transmission method determining unit 14. The interleaving unit 52 sends the interleaved codeword bit sequence to the modulation unit 17.

  The deinterleaving unit 51 performs a deinterleaving process on the bit likelihood sequence output from the demodulating unit 12 using a deinterleaving pattern opposite to the interleaving pattern used in the interleaving unit of the communication device that is the transmission source of the received signal. Undo the interleaved sequence. The deinterleaving unit 51 sends the deinterleaved bit likelihood sequence to the decoding unit 13.

  A receiving device 3A shown in FIG. 13 performs power line communication with the communication device 2A shown in FIG. 12, and has a configuration in which a deinterleave unit 56 is added to the communication device 3 of FIG. The deinterleaving unit 56 performs a deinterleaving process on the bit likelihood sequence output from the demodulation unit 12 using a deinterleaving pattern opposite to the interleaving pattern used in the interleaving unit 52 of the communication device 2A on the transmission side, Restore the interleaved sequence. The deinterleaving unit 56 sends the deinterleaved bit likelihood sequence to the decoding unit 33. The interleave pattern is notified from the communication device 2A stored in the header to the communication device 2B. The notification method is not limited to this. For example, the interleave pattern may be notified from the communication device 2A to the communication device 2B using a signal different from the transmission signal.

  (4) In the above example, the tone map has been described with respect to the case where BPSK is specified for all subcarriers. However, the present invention is not limited to this, and multi-level PSK (Phase Shift Keying) or QAM is used. (Quadrature Amplitude Modulation), PAM (Pulse Amplitude Modulation) may be used, or a different modulation method may be used for each subcarrier. In this case, the reference for estimating the number of erasures, the length of the erasure pattern, and the position of the erasure bit from the noise generation condition differs depending on the modulation method. For example, when QPSK is used, since 2 bits are transmitted by 1 subcarrier, 2 bits may be lost simultaneously when the subcarrier is affected by external noise. The erasure pattern estimation unit 21 estimates the erasure pattern in consideration of the modulation scheme (number of transmission bits) of each subcarrier used at the time of transmission.

  (5) The communication device 2 of the first embodiment observes the noise generation state during non-communication, and displays all the code lengths, coding rates, interleave patterns, and tone maps that may be specified by the control signal. On the other hand, a parity check matrix that obtains a predetermined decoding performance by estimating the erasure pattern and decoding performance from the observed noise generation situation is adopted and used at the time of data transmission, but it is not limited to this configuration. Absent. For example, during non-communication, the noise generation status is observed, and at the time of data transmission, the erasure pattern is determined from the noise generation status according to the code length specified by the control signal, its coding rate, interleave pattern, and tone map. A parity check matrix having good decoding performance may be determined. By doing so, a parity check matrix that can obtain a predetermined decoding performance for all code lengths, coding rates, interleave patterns, and tone maps that may be specified by the control signal during non-communication. It is only necessary to obtain a parity check matrix that provides a predetermined decoding performance under the conditions used for data transmission. For example, the same modification can be applied to the mode for determining the interleave pattern of the codeword bit sequence described in (3) above.

  (6) In communication apparatus 2 of the first embodiment, basic parity check matrix column replacement or codeword bit sequence so that decoding performance can be improved under the condition of the erasure pattern estimated by erasure pattern estimation section 21. Although the configuration for performing rearrangement (for example, the description of (1-3) above applies) has been described, the configuration is not limited to this configuration. For example, a parity check matrix that recommends a parity check matrix that provides the best decoding performance may be configured from a plurality of parity check matrices that have the same code length and coding rate and are independent of each other. Here, the relationship between matrices that are not the same even when column replacement, row replacement, and addition processing of a plurality of rows are performed is called independent. In this case, the communication apparatus 2 needs to hold a plurality of independent parity check matrices. However, when the deterioration of the decoding performance cannot be avoided only by column replacement, the communication apparatus 2 can perform decoding by using the independent parity check matrix. You can go around improving the performance.

(7) In the first embodiment, using the fact that the external noise observed in the communication device 2 and the communication device 3 has a correlation, the communication device 2 multiplies the output signal of the receiving unit 11 during non-communication. An erasure pattern is estimated from the noise level of each subcarrier obtained by carrier demodulation, and the erasure pattern estimated by the own apparatus is used to estimate decoding performance.

  In the following, regarding a communication system that feeds back the noise generation status in the receiving communication device to the transmitting communication device when the correlation between the external noises observed in the transmitting communication device and the receiving communication device is low explain.

  FIG. 14 is a system configuration diagram of a communication system according to a modification of the first embodiment. The communication system 1B in FIG. 14 shows an example in which the communication device 2B connected to the power line 5A and the communication device 3B connected to the power line 5B communicate across the distribution board 7. At this time, the noise source 4 of the external noise exists in the same system as the communication device 2B. In power line communication, it is known that the signal level greatly attenuates the transmission via the distribution board. Further, when a high frequency band is used for signal transmission of power line communication, the transmission path characteristic from the noise source to the communication apparatus differs for each communication apparatus. Therefore, the transmission side communication apparatus 2B observes due to the influence of the transmission path characteristic. The extraneous noise that could not be made may greatly affect the communication device 3B on the receiving side. In this case, the correlation between the external noises observed by the communication device 2B and the communication device 3B is low. As a result, the erasure pattern indicating the arrangement of erasure candidate bits estimated by the communication apparatus 2B is different from the erasure pattern indicating the arrangement of bits lost when the communication apparatus 3B performs decoding, and is observed by the communication apparatus 2B. In some cases, the parity check matrix control based on extraneous noise cannot improve the decoding performance. In order to solve this problem, the communication device 3B includes a configuration that feeds back the erasure pattern used for decoding to the communication device 2B when the decoding result is not good.

  15 is a diagram showing the configuration of the communication device 2B of FIG. 14, and FIG. 16 is a diagram showing the configuration of the communication device 3B of FIG.

  Compared to the communication device 2 shown in FIGS. 9 and 10, the communication device 2 </ b> B shown in FIG. 15 replaces the reception characteristic estimation unit 14 with a reception performance estimation unit 14 </ b> B provided with a decoding performance estimation unit 22 </ b> B instead of the decoding performance estimation unit 22. It has a replaced configuration. The decoding performance estimator 22B uses the erasure pattern (the number of codeword bits lost by the demodulator 12B and the codeword lost) corresponding to the bits lost by the demodulator 12B (to be described later) in the communication device 3B. Until the erasure pattern indicating the position of each bit in the codeword is fed back, the erasure pattern sent from the erasure pattern estimation unit 21 is used to obtain good decoding performance (predetermined decoding performance is obtained). ) Obtain the parity check matrix and send the recommended transmission method including the obtained parity check matrix to the transmission method determination unit 15. In addition, the decoding performance estimation unit 22B receives the erasure pattern received from the communication device 2B as feedback information after receiving the erasure pattern corresponding to the bit lost by the demodulation unit 12B described later in the communication device 3B from the communication device 3B. A parity check matrix having good decoding performance (a predetermined decoding performance is obtained) is obtained and a recommended transmission method including the obtained parity check matrix is sent to the transmission method determination unit 15.

The communication device 3B shown in FIG. 16 has a configuration in which the demodulation unit 12 is replaced with a demodulation unit 12B, as compared with the communication device 2 of the first embodiment. The demodulation unit 12B performs multicarrier demodulation processing on the reception signal transmitted from the reception unit 11, and designates the modulation method of each subcarrier of the multicarrier modulation signal used in the communication device that is the transmission source of the reception signal. The subcarrier is demodulated according to the tone map to calculate the bit likelihood of each transmitted codeword bit. Here, as described above, the demodulation unit 12B erases the code word bit determined to be strongly affected by the external noise, with the bit likelihood being 0. The demodulator 12B sends the bit likelihood sequence to the decoder 13. At the same time, the demodulator 12B creates an erasure pattern corresponding to the lost codeword bits (the number of lost codeword bits and the erasure pattern representing the positions of the lost codeword bits in the respective codewords). Then, the created disappearance pattern is output.

  The communication device 3B holds the erasure pattern output from the demodulation unit 12B until the next time a signal is transmitted to the communication device 2B. When the communication device 3B transmits, for example, an acknowledgment signal (ACK), a non-acknowledge response signal (NACK), or transmission data addressed to the communication device 2B, the demodulation unit 12B outputs the signal when receiving the signal from the communication device 2B. An erasure pattern is included in these signals and transmitted.

  The communication device 2B receives data including the erasure pattern from the communication device 3B, and obtains the erasure pattern by the processing of the coupling circuit 10, the reception unit 11, the demodulation unit 12, and the decoding unit 13. The obtained erasure pattern is input to the decoding performance estimation unit 22B in the reception characteristic estimation unit 14B in order to determine a parity check matrix to be used at the next communication with the communication device 3B. Then, the decoding performance estimation unit 22B obtains a parity check matrix that can obtain a predetermined decoding performance using the input erasure pattern (the erasure pattern transmitted from the communication device 3B).

  When communicating with the communication device 3B, the communication device 2B determines a parity check matrix to be used for encoding transmission data based on the processing result of the decoding performance estimation unit 22B, performs communication such as LDPC encoding of the transmission data, and the like. A signal is transmitted to the device 3B.

  This realizes transmission with good decoding performance in the reception device 3B even when the correlation between the external noise observed in the communication device 2B and the external noise observed in the communication device 3B is low. be able to.

  Note that, for example, a form in which codeword bits in a codeword of a codeword bit sequence are replaced (for example, the above (1-3) corresponds), a form in which a codeword bit sequence is interleaved (for example, the above-mentioned (3) also applies) to the configuration in which the erasure pattern output from the demodulator of the communication device on the receiving side is fed back.

  Further, the information transmitted from the communication device 3B to the communication device 2B may not be the disappearance pattern itself. For example, a parity check matrix having good decoding performance (a predetermined decoding performance can be obtained) is obtained from the erasure pattern output by the demodulation unit 12B in the decoding performance estimation unit in the reception characteristic estimation unit of the communication device 3B. Then, the decoding performance estimation unit determines a transmission method including the obtained parity check matrix, the code length of the LDPC code used for obtaining the parity check matrix, the coding rate, the interleave pattern of the codeword bit sequence, and the tone map. The recommended transmission method may be used, and information related to the recommended transmission method may be included in the transmission data and transmitted to the communication device 2B. When the communication device 2B receives the transmission data including the recommended transmission method, the next transmission to the communication device 3B includes the recommended transmission method in the control signal and the recommended transmission method specified by the control signal. A signal instructing transmission processing with the included information is input to the transmission method determination unit. The transmission method determination unit determines the recommended transmission method indicated by this signal as the transmission method, and issues an instruction to the encoding unit 15 and the modulation unit 16 to perform encoding and modulation using the determined transmission method.

<< Embodiment 2 >>
In the first embodiment, the configuration for improving the decoding performance by changing the parity check matrix used for LDPC encoding, the interleave pattern of the codeword bit sequence, and the like has been described. However, in the second embodiment, the multicarrier is used. A configuration for improving the decoding performance by changing the tone map for designating the modulation method of each subcarrier of the modulation signal will be described. In the second embodiment, components that are substantially the same as those in the first embodiment are denoted by the same reference numerals, and the description thereof can be applied. Therefore, the description thereof is omitted in the second embodiment.

As described above, in an exogenous noise environment, the influence of the exogenous noise at the time of decoding can be reduced by decoding with the bit likelihood of a codeword bit that is strongly influenced by the exogenous noise being 0.

  In the first embodiment, the case where BPSK is used for all subcarriers of a multicarrier modulation signal has been described as a specific example. However, in this case, the number of codeword bits transmitted by one subcarrier is one, so There is one erasure bit per subcarrier affected by noise.

  On the other hand, when a multilevel modulation scheme such as QPSK or QAM is used, the codeword bits transmitted by one subcarrier is 2 bits or more, so when there are subcarriers affected by external noise, Two or more erasure bits are generated per subcarrier. Therefore, when a subcarrier that is specified to use a modulation scheme with a high modulation multi-level number is subjected to extraneous noise by the tone map, and a subcarrier that is specified to use a modulation scheme with a low modulation multi-level number is The number of erasure bits changes when affected by extraneous noise. As the number of lost bits increases, the decoding performance tends to deteriorate. In particular, when the number of erasure bits is close to the number of parity bits of the LDPC code used, the decoding performance is greatly degraded.

  The communication apparatus according to Embodiment 2 observes the occurrence of extraneous noise, and uses a tone map that uses a low modulation multi-level modulation scheme for subcarriers that are strongly affected by extraneous noise. This is intended to improve the decoding performance under the environment.

  FIG. 17 is a diagram illustrating a configuration of a communication device 2D having a function of changing a tone map in the second embodiment. Compared to the communication device 2, the communication device 2D has a configuration in which the reception characteristic estimation unit 14 and the transmission method determination unit 15 are replaced with a reception characteristic estimation unit 14D and a transmission method determination unit 15D.

  As shown in FIG. 18, the reception characteristic estimation unit 14D includes a noise generation state estimation unit 20, an erasure pattern estimation unit 21D, and a decoding performance estimation unit 22D.

  The erasure pattern estimation unit 21D uses the information on the subcarriers affected by the external noise transmitted from the noise generation state estimation unit 20 and the information on the number of bits to be transmitted in one symbol of the multicarrier modulation signal, as erasure candidates. Create a tone map with fewer bits. Then, the erasure pattern estimation unit 21D estimates the erasure pattern when the created tone map is used. However, the erasure pattern estimation unit 21D has a subcarrier number that is affected by extraneous noise, a tone map that specifies the modulation scheme of each subcarrier, the code length and coding rate of the LDPC code, and the codeword bit sequence. The disappearance pattern is estimated from the interleave pattern. The erasure pattern estimation unit 21D sends to the decoding performance estimation unit 22D a tone map in which the number of erasure candidate bits is reduced and an erasure pattern estimated using the tone map.

  The decoding performance estimation unit 22D estimates the decoding performance when decoding is performed using the erasure pattern sent from the erasure pattern estimation unit 21D. If the estimated decoding performance is good, the decoding performance estimation unit 22D, the tone map that is the basis for creating the erasure pattern, the code length and coding rate of the LDPC code, the interleave pattern of the codeword bit sequence, and the decoding performance The recommended transmission scheme including the parity check matrix used for estimation of the transmission scheme is sent to the transmission scheme determination unit 15D.

The transmission method determination unit 15D transmits tones received from the reception characteristic estimation unit 14D for the interleave patterns of the code lengths, coding rates, and codeword bit sequences of all LDPC codes that may be specified by the control signal. Based on the recommended transmission method including maps, etc., and the contents of instructions in the input control signal (LDPC code length, coding rate, codeword bit sequence interleaving pattern), the transmission data is actually transmitted. Determine the transmission method including the tone map to be used. Then, the transmission method determination unit 15 instructs the encoding unit 16 to perform LDPC encoding processing of transmission data using D, the code length of the LDPC code included in the determined transmission method, the coding rate, and the parity check matrix. . In addition, the transmission scheme determination unit 15D instructs the modulation unit 17 to modulate each subcarrier according to the tone map included in the determined transmission scheme.

  Hereinafter, the operation of the communication device 2D at the noise level shown in FIG. 19 will be described. Here, it is assumed that communication apparatus 2D transmits 20 codeword bits in one multicarrier modulation signal symbol. At this time, if the parity check matrix H shown in Equation (1) is used, since the code length of the parity check matrix H is 10, two codewords (codeword 1 and codeword 2) are converted into one multicarrier modulation. The signal symbol is transmitted.

  The noise generation state estimation unit 20 in the reception characteristic estimation unit 14D compares the noise level in each subcarrier indicated by the noise state estimation signal with the noise threshold Nth. The noise generation state estimation unit 20 derives the subcarrier numbers whose noise level is higher than the noise threshold Nth as fsc1, fsc2,. In the example of FIG. 19, assuming that the subcarrier number is given as 1, 2,..., 10 in order from the lowest frequency in the total number of subcarriers 10, fsc1 = 2, fsc2 = 4, fsc3 = 8, fsc4 = 10 The noise generation situation estimation unit 20 uses the information of the subcarrier number whose noise level is higher than the noise threshold Nth (in the example of FIG. 19, fsc1 = 2, fsc2 = 4, fsc3 = 8, fsc4 = 10) as the noise occurrence situation and disappears. This is sent to the pattern estimation unit 21D.

  In order to improve the decoding performance when there is an erasure, the number of codeword bits transmitted on subcarriers having a subcarrier number higher than the noise threshold Nth may be reduced. Therefore, the erasure pattern estimation unit 21D uses BPSK for the second, fourth, eighth, and tenth subcarriers, uses QPSK for the first, third, seventh, and ninth subcarriers, and uses the fifth and sixth subcarriers. Then, a tone map using 16QAM is created, and the erasure pattern of each codeword is estimated.

  If the codeword bit sequence is not interleaved, the first codeword bit of codeword 1, the second codeword bit of codeword 1,..., The 10th codeword bit of codeword 1, and the codeword The first codeword bit of 2, the second codeword bit of codeword 2,..., The tenth codeword bit of codeword 2, and in order from the subcarrier with the smallest number to the created tone map When assigned, the erasure pattern of codeword 1 is as shown in equation (18), and the erasure pattern of codeword 2 is as shown in equation (19).

消失パターン推定部２１Ｄが作成した上記のトーンマップを用いることで、消失ビットの数を例えば全サブキャリアでＱＰＳＫを用いる場合（消失ビット数８）に比べて少なくすることができる。

By using the above tone map created by the erasure pattern estimation unit 21D, the number of erasure bits can be reduced as compared with, for example, the case where QPSK is used in all subcarriers (number of erasure bits 8).

  The erasure pattern estimation unit 21D generates the created tone map and the erasure pattern of each codeword (the erasure pattern of codeword 1 shown in Equation (18), the erasure pattern of codeword 2 shown in Equation (19)). The data is sent to the decoding performance estimation unit 22D.

  The decoding performance estimation unit 22D estimates the decoding performance for each erasure pattern.

  When using the parity check matrix H of Equation (1), the decoding performance estimation unit 22D transmits the codeword 1 of Equation (18) sent from the erasure pattern estimation unit 21D from the parity check matrix H of Equation (1). Only the columns (3 columns, 6 columns) corresponding to the codeword bits having a value of 1 in the erasure pattern are extracted, and a partial matrix Hp1 is created from the extracted columns. The created submatrix Hp1 is as shown in Equation (20).

復号性能推定部２２Ｄは、式（１）のパリティ検査行列Ｈから、消失パターン推定部２１Ｄから送られてくる式（１９）の符号語２の消失パターンにおいて値が１となっている符号語ビットに対応する列（７列、１０列）のみを抽出して、抽出した列から部分行列Ｈｐ２を作成する。作成される部分行列Ｈｐ２は式（２１）のようになる。

The decoding performance estimation unit 22D obtains the codeword bit having a value of 1 in the erasure pattern of the codeword 2 of the equation (19) sent from the erasure pattern estimation unit 21D from the parity check matrix H of the equation (1). Only the columns corresponding to (7 columns, 10 columns) are extracted, and a partial matrix Hp2 is created from the extracted columns. The created submatrix Hp2 is as shown in Equation (21).

式（２０）に示す部分行列Ｈｐ１、及び式（２１）に示す部分行列Ｈｐ２は、行重みが１の行が１行もなく、行重みが２の行が２つであり、大きな復号性能の劣化はない（所定の復号性能が得られる）。

In the submatrix Hp1 shown in the equation (20) and the submatrix Hp2 shown in the equation (21), there is not one row having a row weight of 1, and two rows having a row weight of 2, which has a large decoding performance. There is no degradation (predetermined decoding performance is obtained).

  The decoding performance estimation unit 22D sends the recommended transmission method including the tone map that is the basis of the erasure pattern of the equations (18) and (19) to the transmission method determination unit 15D.

  As described above, the communication device 2D according to the second embodiment performs transmission using a tone map in which the modulation level of a subcarrier that is highly influenced by external noise is low, and thus, under an external noise environment. Degradation of the decoding performance in the transmission can be reduced. Note that a communication device having the configuration of FIG. 11 can be used as a communication device that performs power line communication with the communication device 2D.

  In the above description, the submatrix given by Equation (20) and Equation (21) has two rows with a row weight of 2, so there is no significant degradation in decoding performance. A configuration in which codewords are interleaved so as to be obtained may be adopted. For example, when codeword interleaving is performed so that the erasure pattern of codeword 1 given by equation (18) becomes equation (22), submatrix Hp3 becomes equation (23).

部分行列Ｈｐ３は、重み２の行が１行のみであるため、この部分行列が得られるトーンマップとインタリーブパターンを含む伝送方式を推奨伝送方式として伝送方式決定部１５Ｄに送る。また、符号語２の消失パターンについても、同様である。

Since the partial matrix Hp3 has only one row of weight 2, the transmission method including the tone map and the interleave pattern from which the partial matrix is obtained is sent to the transmission method determination unit 15D as the recommended transmission method. The same applies to the disappearance pattern of codeword 2.

  Note that both the submatrix Hp1 and the submatrix Hp3 can suppress significant deterioration in that they do not constitute a stopping set, but the decoding performance when erasure bits are included is always better than the submatrix Hp3. Not necessarily. Therefore, for a plurality of transmission methods in which each erasure pattern does not constitute a stopping set such as the partial matrix Hp1 and the partial matrix Hp3, the decoding performance is confirmed in advance by simulation or the like, and the decoding performance estimation unit 22D includes the information Based on this, it may be configured to determine a transmission method with high decoding performance and notify the transmission method determination unit 15D as a recommended transmission method.

In the second embodiment, the erasure pattern estimation unit 21D employs a configuration in which the tone map in which the number of erasure candidate bits is reduced and the erasure pattern when the tone map is used are estimated. It is not limited. For example, the erasure pattern estimation unit 21D estimates the code length and the coding rate in addition to the tone map in which the number of erasure candidate bits is reduced, and uses the tone map, the code length, and the coding rate as the erasure pattern. A configuration of estimating may be used. In this way, not only the tone map but also the type of code such as code length and coding rate can be changed, so that the degree of freedom in estimating a transmission method with little deterioration in decoding performance is improved.

  In the second embodiment, a configuration has been described in which the erasure pattern estimation unit 21D creates a tone map so that BPSK is used for all subcarriers that are estimated to be affected by extraneous noise. It is not something that can be done.

  For example, the erasure pattern estimation unit 21D determines that the modulation multilevel number of the subcarrier modulation scheme affected by strong external noise is the modulation multilevel number of the subcarrier modulation scheme not affected by strong external noise. A tone map may be created so as to reduce the number.

  Also, for example, the erasure pattern estimation unit 21D creates an erasure pattern when a multilevel modulation scheme such as QPSK is used for a part of subcarriers estimated to be affected by extraneous noise, and decoding performance estimation As a result of the determination of the decoding performance for the erasure pattern by the unit 22D, if the deterioration of the decoding performance is acceptable, a multivalue such as QPSK is estimated for some of the subcarriers estimated to be affected by extraneous noise. You may comprise so that a modulation system may be used. This increases the number of bits transmitted in one symbol and improves transmission efficiency.

  In this way, when a multi-level modulation scheme such as QPSK is used with subcarriers estimated to be affected by extraneous noise, the number of erasure candidate bits that can tolerate degradation in decoding performance is calculated in advance and set as a threshold value. If the erasure pattern estimation unit 21D is affected by external noise within a range in which the number of erasure candidate bits in one codeword or the number of erasure candidate bits per symbol does not exceed the threshold value. The estimated subcarrier may be configured to estimate a tone map using a multi-level modulation method such as QPSK.

  In the second embodiment, when erasure pattern estimation unit 21D uses BPSK for subcarriers estimated to be affected by extraneous noise, 16QAM is used for at least some subcarriers that are less affected by extraneous noise. As an example, a configuration in which the number of codeword bits transmitted in one symbol is N times the code length (N is an arbitrary natural number, N = 2 in the second embodiment) is used as an example. Explains. According to this configuration, when transmitting a plurality of symbols in succession, N codewords are arranged and transmitted on subcarriers according to the same arrangement rule for each symbol, so the transmission method used is the same. If there are, the estimated erasure pattern is the same N types of erasure patterns for any codeword transmitted in any symbol (in the case of the second embodiment, the two erasure patterns shown in equations (18) and (19)). Therefore, the reception characteristic estimation unit 14D can determine the recommended transmission scheme in consideration of only the estimated N types of erasure patterns, and the recommended transmission scheme can be easily estimated. In this case, the reception characteristic estimation unit 14D uses, as the transmission method determination unit 15D, one transmission method with little deterioration in decoding performance for any of the N types of erasure patterns as the recommended transmission method, as in the second embodiment. As a recommended transmission method, a parity check matrix with little degradation in decoding performance is estimated for each of N erasure patterns and notified to the transmission method determination unit 15D. The parity check matrix may be switched.

Embodiment 2 is not limited to a configuration in which the number of codeword bits transmitted in one symbol is N times the code length. For example, the number of codeword bits transmitted in one symbol may be configured to be 1 / M times the code length (M is an arbitrary natural number). In this case, since one codeword is arranged in each subcarrier of M symbols with the same arrangement rule, the estimated erasure pattern for a codeword transmitted with any M symbols is It will always be the same. According to this configuration, the reception characteristic estimation unit 14D estimates one transmission method with little deterioration in decoding performance with respect to the estimated erasure pattern, and notifies the transmission method determination unit 15D of the estimated transmission method as a recommended transmission method. Since it is possible to suppress degradation in decoding performance for all codewords, it is easy to estimate a recommended transmission method.

  Further, the second embodiment may have a configuration in which the number of codeword bits transmitted in one symbol and the code length are not divisible by one. In this case, since the arrangement rules for allocating codewords to subcarriers are different for each symbol, the erasure pattern changes for each codeword even if the same transmission method is used, but the erasure pattern of each codeword is a tone map, and Since the calculation can be performed based on the interleave pattern, the reception characteristic estimation unit 14D may determine the recommended transmission method in consideration of a plurality of types of erasure patterns estimated for each codeword to be transmitted. In this case, the reception characteristic estimation unit 14D may notify the transmission method determination unit 15D of one recommended transmission method with little degradation in decoding performance as a recommended transmission method for all estimated erasure patterns, or recommend it. As a transmission method, a transmission method with little deterioration in decoding performance may be estimated for each codeword and notified to the transmission method determination unit 14D, and the transmission method determination unit 14D may switch the transmission method for each codeword.

  In Embodiment 1 and Embodiment 2 described above, the configuration using a modulated signal that has been subjected to multicarrier modulation as a transmission signal has been described as an example. However, the present invention is applicable to a communication apparatus that uses single carrier modulation. You can also. For example, in an environment where strong extraneous noise is generated at a specific time interval, it is expected that erasure candidate bits are arranged at a period corresponding to the occurrence interval of extraneous noise. A combination of codeword bits arranged at a period corresponding to the generation interval of extraneous noise can be estimated as an erasure pattern. In this case, there are multiple combinations of codeword bits arranged at a period corresponding to the generation interval of the external noise, and which erasure pattern is determined is the timing at which the external noise is generated and the timing at which transmission of the modulation signal is started It is expected to vary depending on the relationship between For this reason, the reception characteristic estimation unit 14 and the like estimate a transmission method in which deterioration of decoding performance is reduced in any of a plurality of erasure patterns, and notify the transmission method determination unit 15 and the like as the recommended transmission method. It ’s fine.

  In the first embodiment, the erasure pattern is changed by changing the parity check matrix, the interleave pattern of the codeword bit sequence, and the like, and the decoding performance is improved. By changing the tone map in which the modulation scheme of subcarriers affected by noise is BPSK with the smallest number of modulation multilevels, the number of erasure candidate bits is reduced to improve the decoding performance. However, it is not limited to this, For example, the following may be sufficient. If the tone map is changed, the erasure pattern is changed, and even if the number of erasure candidate bits does not change or the number of erasure candidate bits increases, depending on the combination of erasure candidate bit positions in the codeword Good decoding performance may be obtained. Therefore, the erasure pattern is estimated for each of a plurality of tone maps (for example, the number of subcarriers specifying each modulation method is the same among the plurality of tone maps), and the decoding performance is estimated. Alternatively, a tone map that uses a tone map with good decoding performance may be determined.

  In the first embodiment and the second embodiment, the power line communication system indicated by the communication systems 1 and 1C has been described, but the present invention can also be applied to a wireless communication system such as a wireless LAN.

In addition, each functional block included in the communication device (the communication device on the transmission side and the communication device on the reception side) described in the first and second embodiments and the modifications thereof is typically an integrated circuit. This is realized as an LSI. These may be individually made into one chip, or may be made into one chip so as to include a part or all of them. The name used here is LSI, but it may also be called IC, system LSI, super LSI, or ultra LSI depending on the degree of integration. Further, the integration method is not limited to LSI, and may be realized by a dedicated circuit or a general-purpose processor. An FPGA (Field Programmable Gate Array) that can be programmed after manufacturing the LSI or a reconfigurable processor that can reconfigure the connection and setting of circuit cells inside the LSI may be used. Furthermore, if integrated circuit technology comes out to replace LSI's as a result of the advancement of semiconductor technology or a derivative other technology, it is naturally also possible to carry out function block integration using this technology. There is a possibility of adaptation of biotechnology.

  The communication device, the reception device, the transmission method, and the reception method according to the present invention are useful as a communication device for digital communication under an external noise environment.

DESCRIPTION OF SYMBOLS 1 Communication system 2 Communication apparatus 3 Communication apparatus 4 Noise source 5 Power line 10 Coupling circuit 11 Reception part 12 Demodulation part 13 Decoding part 14 Reception characteristic estimation part 15 Transmission system determination part 16 Encoding part 17 Modulation part 18 Transmission part 20 Noise generation Situation estimation unit 21 Erasure pattern estimation unit 22 Decoding performance estimation unit 30 Coupling circuit 31 Reception unit 32 Demodulation unit 33 Decoding unit

Claims (4)

A transmission device that transmits transmission data using a transmission method selected from a plurality of transmission methods including at least a first transmission method and a second transmission method,
A first codeword bit sequence is generated by encoding the transmission data using an LDPC encoding method selected from a plurality of LDPC encoding methods having different code lengths and encoding rates from each other. An encoding unit;
An interleaving unit that generates a second codeword bit sequence by interleaving the first codeword bit sequence using an interleave pattern selected from a plurality of different interleave patterns;
A modulation unit that generates a modulation symbol by performing mapping on a predetermined number of bits included in the second codeword bit sequence using a modulation method selected from a plurality of different modulation methods;
A control symbol generated from control information including information indicating an LDPC encoding scheme, an interleave pattern, and a modulation scheme used for transmission of the transmission data, and a transmitter that transmits the modulation symbol,
The first transmission scheme is a scheme using a first LDPC coding scheme, a first interleave pattern, and a first modulation scheme, and the second transmission scheme is a first LDPC coding scheme and a second modulation scheme. Using the first interleaving pattern and the first modulation scheme, the first interleaving pattern and the second interleaving pattern being different from each other,
Transmitter device. A transmission method for transmitting transmission data using a transmission method selected from a plurality of transmission methods including at least a first transmission method and a second transmission method,
A first codeword bit sequence is generated by encoding the transmission data using an LDPC encoding method selected from a plurality of LDPC encoding methods having at least one of a code length and a coding rate different from each other. ,
Interleaving the first codeword bit sequence using an interleave pattern selected from a plurality of different interleave patterns to generate a second codeword bit sequence;
Mapping a predetermined number of bits included in the second codeword bit sequence using a modulation scheme selected from a plurality of different modulation schemes to generate modulation symbols;
Processing for transmitting a control symbol generated from control information including information indicating an LDPC encoding scheme, an interleave pattern, and a modulation scheme used for transmission of the transmission data, and the modulation symbol,
The first transmission scheme is a scheme using a first LDPC coding scheme, a first interleave pattern, and a first modulation scheme, and the second transmission scheme is a first LDPC coding scheme and a second modulation scheme. Using the first interleaving pattern and the first modulation scheme, the first interleaving pattern and the second interleaving pattern being different from each other,
Transmission method. A receiving apparatus that receives a control symbol and a modulation symbol transmitted using a transmission method selected from a plurality of transmission methods including at least a first transmission method and a second transmission method,
The modulation symbol is generated by mapping a predetermined number of bits included in the second codeword bit sequence using a modulation method selected from a plurality of different modulation methods,
The second encoded bit sequence is generated by interleaving the first codeword bit sequence using an interleave pattern selected from a plurality of different interleave patterns,
The first codeword bit sequence is generated by encoding transmission data using an LDPC encoding method selected from among a plurality of LDPC encoding methods having different code lengths and encoding rates. It has been
The control symbol is generated from control information including information indicating an LDPC encoding scheme, an interleave pattern, and a modulation scheme used for transmission of the transmission data,
The first transmission scheme is a scheme using a first LDPC coding scheme, a first interleave pattern, and a first modulation scheme, and the second transmission scheme is a first LDPC coding scheme and a second modulation scheme. The interleave pattern and the first modulation scheme are used, and the first interleave pattern and the second interleave pattern are different from each other,
A demodulator that demodulates the modulation symbol based on the information;
Receiver device. A receiving method for receiving a control symbol and a modulation symbol transmitted using a transmission method selected from a plurality of transmission methods including at least a first transmission method and a second transmission method,
The modulation symbol is generated by mapping a predetermined number of bits included in the second codeword bit sequence using a modulation method selected from a plurality of different modulation methods,
The second encoded bit sequence is generated by interleaving the first codeword bit sequence using an interleave pattern selected from a plurality of different interleave patterns,
The first codeword bit sequence is generated by encoding transmission data using an LDPC encoding method selected from among a plurality of LDPC encoding methods having different code lengths and encoding rates. It has been
The control symbol is generated from control information including information indicating an LDPC encoding scheme, an interleave pattern, and a modulation scheme used for transmission of the transmission data,
The first transmission scheme is a scheme using a first LDPC coding scheme, a first interleave pattern, and a first modulation scheme, and the second transmission scheme is a first LDPC coding scheme and a second modulation scheme. The interleave pattern and the first modulation scheme are used, and the first interleave pattern and the second interleave pattern are different from each other,
Based on the information, the modulation symbols are demodulated to generate received data.
Reception method.

Images